Robert Reich: Stop AI Before It’s Too Late – OpEd
August 12, 2026
By Robert Reich
Key Takeaways
Recent U.S. job losses (23,000 in July) and slowing wage growth, especially in AI-exposed occupations, indicate that artificial intelligence is already contributing to higher unemployment and reduced earnings for millions of workers.
AI is concentrating vast wealth and political power among a small group of executives and investors while imposing heavy environmental costs through energy-intensive data centers and raising serious safety risks, including rogue models and potential bioweapon development.
Given that the known costs and risks currently outweigh broad benefits for most people, the rapid advance of AI should not be treated as inevitable; society has both the right and the responsibility to pause or halt its development to protect jobs, democracy, the climate, and public safety.
I’m going to make a proposal today that’s almost certain to get me consigned to the neo-Luddite dustbin of history.
But first, let me lay out some facts.
Rather than producing jobs, the U.S. economy actually lost 23,000 job in July, according to Bureau of Labor Statistics data released Friday. In addition, May’s and June’s job numbers were revised downward, showing a combined 103,000 fewer jobs than previously reported.
As if this weren’t bad enough, wage growth has also slowed. Average hourly earnings In July were just 0.1 percent higher than in June. This isn’t just a single month’s slow wage growth, either. Average hourly earnings increased just 3.2 percent over the past year — the lowest annual growth rate in five years.
What’s going on? It’s too early to tell. But evidence is mounting that artificial intelligence is playing a role.
New research by economists at Morgan Stanley shows that the rate of unemployment is half a percentage point higher than it would otherwise be in occupations exposed to AI, which they put at about 30 percent of all employment. The effect is even more dramatic among younger people.
Wage growth in jobs exposed to AI has contracted by 6.7 percent since 2023, according to research by economists Sania Edlich and Apollo Global Management’s Torsten Slok. This has resulted in at least $28 billion in losses for 5.8 million affected workers.
These findings still don’t explain the startling loss of jobs in July or the downward revisions for May and June. There are probably many factors at play. But they suggest that employers may be anticipating they’ll need fewer workers in the future — and won’t need to pay them all that much in order to attract them.
It’s possible that AI may create more jobs over the long term. But as John Maynard Keynes once noted, over the long term we’re all dead.
More than half of Americans surveyed by Reuters/Ipsos in June say they’re worried AI will put someone in their household out of work.
Edlich and Slok write that “the critical policy question is not whether AI will reshape the labor market more broadly, but how quickly, and whether workers will have the support they need when it does.”
As a former secretary of labor who’s kept his eyes focused on the Trump regime, I can assure you workers won’t have the support they need any time soon.
And even if AI begins to generate the productivity bonanza its advocates predict — but hasn’t yet — there’s no reason to assume American workers will see any of the benefits in their paychecks. If you hadn’t noticed, wages have been stuck even as the stock market has roared.
To the contrary, all signs point to vast riches for a few major AI investors and executives while most Americans are left behind.
Wealth inequality is already at record levels, and wealth at the top is quickly morphing into political power.
AI is creating a vast wave of campaign money. OpenAI’s superPAC “Leading the Future” has amassed over $140 million to influence upcoming elections, while Anthropic’s superPAC “Public First Action” isn’t far behind.
As the great jurist Louis Brandeis is reputed to have said, “America has a choice: we can have great wealth in the hands of a few, or we can have a democracy, but we can’t have both.”
AI is pushing us further toward the first option.
Meanwhile, there’s the planet to consider.
Amazon is now investing in a large-scale natural-gas power plant as part of a huge data center in Pecos County, Texas — a facility that could become the largest single source of climate pollution in the United States.
It’s racing to build enough data centers to keep pace with other giant AI corporations and secure the electricity to power them. Amazon’s new gas-burning plant is permitted to release 33 million tons of carbon dioxide a year, regulatory records show, more planet-warming gases than any other power plant in America.
So much for Amazon’s promise to eliminate its planet-warming emissions by 2040 as part of its Climate Pledge. You can bet other giants in the AI race will be turning to natural gas, too.
Oh, and I haven’t even mentioned the Frankenstein monster in the room. A few weeks ago, OpenAI admitted that two of its artificial intelligence models went rogue and successfully hacked into a digital library of AI technology.
The incident, which happened while OpenAI was testing the cybersecurity capabilities of its systems, was the kind of science-fiction nightmare that could soon be a reality. How soon before AI models escape all their cages?
Just last week, scientists published a study documenting how they used A.I. to create new kinds of viruses, raising the frightful possibility that the technology could be used to invent dangerous pathogens.
Lost jobs. Lost wages. Widening inequality. Data centers using up water and electricity and polluting the climate. Vastly more money polluting our politics. Models escaping their cages and hacking into everything, possibly threatening human life on this planet.
Can we pause for a moment and talk about what’s really happening here?
As sociologist Tressie McMillan Cottom writes, AI has merged regressive politics with unchecked economic power under the guise of technological innovation.
Far too much money is giving a small group of unelected people extraordinary power to determine our future in ways that are likely to remake — and could possibly destroy — our lives.
We’re watching all of this roll out as if we have no choice, as if it’s inevitable, as if AI is just something we’re going to have to adapt to.
But why should we have to adapt to it, when it is the product of people like Jeff Bezos, Elon Musk, Sam Altman, Mark Zuckerberg, and Dario Amodei?
Why should we be confined to being spectators at their enormously dangerous game? Why should we have to accept all these hugely negative, potentially life-threatening consequences?
The fact is, we don’t.
Communities across America are organizing against data centers near them. MAGAs and progressives are joining together to say “no” to the noise, higher electricity bills, and water shortages.
Well, then, why can’t we stop the whole damn thing? Why can’t we decide that the incalculable costs and risks of AI aren’t worth the potential benefits to the vast majority of us?
AI proponents argue that stopping or even pausing AI in the United States would risk American industry falling behind competitors overseas.
But if the costs and risks exceed known benefits, why not let China or any other competitor try AI out first? Why should we be the canary in this extraordinarily dangerous coal mine?
Other advocates of AI say we have no right to stop innovation in the free market. That’s baloney. We don’t allow private corporations to come up with new types of nuclear weapons or varieties of cocaine or biological pathogens. We protect the public from certain kinds of innovation.
So let’s protect ourselves here. Stop AI before it’s too late.
This article was published at Robert Reich’s Substack
About Robert Reich
Robert B. Reichis Chancellor's Professor of Public Policy at the University of California at Berkeley and Senior Fellow at the Blum Center for Developing Economies, and writes atrobertreich.substack.com. Reich served as Secretary of Labor in the Clinton administration, for which Time Magazine named him one of the ten most effective cabinet secretaries of the twentieth century. He has written fifteen books, including the best sellers "Aftershock", "The Work of Nations," and"Beyond Outrage," and, his most recent, "The Common Good," which is available in bookstores now. He is also a founding editor of the American Prospect magazine, chairman of Common Cause, a member of the American Academy of Arts and Sciences, and co-creator of the award-winning documentary, "Inequality For All." He's co-creator of the Netflix original documentary "Saving Capitalism," which is streaming now.
View all posts by Robert Reich →
A Democracy Cannot Run on an AI Model
AI is already inside election administration; before it moves deeper into the count, results have to stay verifiable.
Clerk Matthew Sandbar, 34, demonstrates part of the ballot-sorting process to members of the media attending a walkthrough of the Philadelphia Ballot Processing Center used to process the ballots of the 2022 US midterm elections, in Philadelphia on October 27, 2022.
(Photo by Ryan Collerd / AFP via Getty Images)
Davis Austria
Aug 12, 2026Common Dreams
Tallying the results of Washington, DC’s first ranked-choice election in June took about 10 days—longer than many voters are used to. In an age when artificial intelligence can generate an answer in seconds, waiting days for election results can feel old-fashioned, inefficient, even suspicious. And it is about to matter far more widely. This November, 17 states, cities, and counties will use ranked-choice voting, including Maine and Alaska statewide, in races that could help decide control of Congress. More voters than ever will watch results take days to resolve, and more will be asked to trust a count they cannot see.
But that slowness may be one of its most democratic features.
I come to this question from health informatics, where I study how even AI-generated information that sounds clear, fluent, and helpful still requires careful review before it reaches a patient. The same verification problem applies to elections. The more authoritative a system sounds, the more important it becomes to make sure the output can be checked.
DC’s June 16 primary followed a system that requires more than a simple tally: Voters rank candidates, and if no candidate receives more than 50% of first-choice votes, lower-performing candidates are eliminated, and votes are redistributed according to voters’ next choices. New York City already uses ranked-choice voting in local primary and special elections.
AI may soon be able to produce election results in seconds. That does not mean it should.
There is nothing wrong with ranked-choice voting, but it must be carried out with verifiable results. Because election counting is not just a math problem; it is a trust problem, especially in the US today, where election workers face harassment, routine counting delays are recast as fraud, and many voters already doubt institutions before a single ballot is counted.
In this climate, counting ranked-choice votes in a way that allows for public verification might take longer than people are used to. Every ballot must be tied to a voter-verifiable record. Every round of tabulation must be explainable. Every disputed outcome must be auditable by people who can inspect the evidence themselves.
AI is not yet counting votes, but AI and machine-assisted systems are already touching elections before ballots are counted, including information voters receive, and how their signatures are reviewed. In 2024, X’s Grok chatbot gave users false information about ballot deadlines; after election officials from five states complained, X changed Grok so election-related questions directed users to Vote.gov, the federal government’s official voting information website. In North Carolina, 10 counties piloted automated signature-verification software for absentee-by-mail ballots in 2024. The pilot did not affect whether any ballot was counted, but later reporting found reliability problems: The software failed to match about 11% of signatures, most software-flagged signatures were approved after human review, and technical issues complicated the test. Neither example is the same as artificial intelligence counting votes, but both show how software can shape what voters are told or what happens before a ballot enters the count.
I have seen the problem in my own research. AI systems can cite studies that do not exist, with titles, authors, and journals that look real until you check them. A 2026 Nature analysis warned that hallucinated citations are polluting scientific literature, and OpenAI researchers have warned that some training and evaluation systems reward guessing over acknowledging uncertainty. A system that gives a confident answer may look more useful, even when that confidence is misplaced, than one that says, “I don’t know.”
That matters because elections often turn on ballot conditions that are not immediately machine readable and need human review: stray marks, undervotes, overvotes, damaged ballots, or ambiguous voter intent. In races where the margin between the candidates is narrow, a few disputed ballots can change the result. These questions should not be resolved by a system whose rules, error rates, or decision process the public cannot understand or meaningfully examine. They should be handled through documented procedures, human review, paper records, and public audits.
Human counting is also imperfect. But precisely because humans are fallible, democratic systems have developed safeguards such as paper ballots, bipartisan observation, chain-of-custody rules, cure processes, recounts, and post-election audits. Risk-limiting audits, for example, hand check samples of paper ballots and can trigger a fuller count if the sample does not support the reported result. Colorado was the first state to conduct a statewide risk-limiting audit in 2017; today, all Colorado counties conduct one before results are certified. Other states have since adopted, required, or piloted risk-limiting audits.
The best election systems do not ask voters to trust either a person or a machine. They produce evidence. That is where current policy is behind the curve. States have moved quickly to regulate AI-generated deepfakes in campaign communications, because fake videos and robocalls can mislead voters. But far less attention has been paid to the AI inside election administration itself. That gap should close before AI moves deeper into election administration, especially into decisions that could affect whether a ballot is accepted, cured, rejected, or counted.
Federal and state governments should establish clear standards for AI used in anything election related. The Brennan Center for Justice has called for such safeguards around quality, transparency, consistency, certification, and monitoring. Election offices should have to explain to the public what an AI system does, who checks it, and how voters can verify the evidence. AI systems whose rules, error rates, or decision process the public cannot understand or meaningfully examine should not make final decisions about whether a ballot counts.
Voters have a role, too. Ask your local election officials whether AI tools are being used anywhere in election administration and how those decisions can be audited. Ask how voters will know if the AI setup gets something wrong. And volunteer as a poll worker. You don’t need a background in elections; local officials train you. Democracy needs more people who understand how ballots are handled, checked, and verified.
AI may soon be able to produce election results in seconds. That does not mean it should. The labor of democracy is not a bottleneck in need of streamlining. It is part of the safeguard.
Our work is licensed under Creative Commons (CC BY-NC-ND 3.0). Feel free to republish and share widely.
Davis Austria
Davis Austria is an assistant professor of health informatics conducting AI research at Xavier University of Louisiana, a Hopelab HBCU translational science fellow, and a Public Voices fellow with The OpEd Project.
Full Bio >
August 12, 2026
By Robert Reich
Key Takeaways
AI is concentrating vast wealth and political power among a small group of executives and investors while imposing heavy environmental costs through energy-intensive data centers and raising serious safety risks, including rogue models and potential bioweapon development.
Given that the known costs and risks currently outweigh broad benefits for most people, the rapid advance of AI should not be treated as inevitable; society has both the right and the responsibility to pause or halt its development to protect jobs, democracy, the climate, and public safety.
I’m going to make a proposal today that’s almost certain to get me consigned to the neo-Luddite dustbin of history.
But first, let me lay out some facts.
Rather than producing jobs, the U.S. economy actually lost 23,000 job in July, according to Bureau of Labor Statistics data released Friday. In addition, May’s and June’s job numbers were revised downward, showing a combined 103,000 fewer jobs than previously reported.
As if this weren’t bad enough, wage growth has also slowed. Average hourly earnings In July were just 0.1 percent higher than in June. This isn’t just a single month’s slow wage growth, either. Average hourly earnings increased just 3.2 percent over the past year — the lowest annual growth rate in five years.
What’s going on? It’s too early to tell. But evidence is mounting that artificial intelligence is playing a role.
New research by economists at Morgan Stanley shows that the rate of unemployment is half a percentage point higher than it would otherwise be in occupations exposed to AI, which they put at about 30 percent of all employment. The effect is even more dramatic among younger people.
Wage growth in jobs exposed to AI has contracted by 6.7 percent since 2023, according to research by economists Sania Edlich and Apollo Global Management’s Torsten Slok. This has resulted in at least $28 billion in losses for 5.8 million affected workers.
These findings still don’t explain the startling loss of jobs in July or the downward revisions for May and June. There are probably many factors at play. But they suggest that employers may be anticipating they’ll need fewer workers in the future — and won’t need to pay them all that much in order to attract them.
It’s possible that AI may create more jobs over the long term. But as John Maynard Keynes once noted, over the long term we’re all dead.
More than half of Americans surveyed by Reuters/Ipsos in June say they’re worried AI will put someone in their household out of work.
Edlich and Slok write that “the critical policy question is not whether AI will reshape the labor market more broadly, but how quickly, and whether workers will have the support they need when it does.”
As a former secretary of labor who’s kept his eyes focused on the Trump regime, I can assure you workers won’t have the support they need any time soon.
And even if AI begins to generate the productivity bonanza its advocates predict — but hasn’t yet — there’s no reason to assume American workers will see any of the benefits in their paychecks. If you hadn’t noticed, wages have been stuck even as the stock market has roared.
To the contrary, all signs point to vast riches for a few major AI investors and executives while most Americans are left behind.
Wealth inequality is already at record levels, and wealth at the top is quickly morphing into political power.
AI is creating a vast wave of campaign money. OpenAI’s superPAC “Leading the Future” has amassed over $140 million to influence upcoming elections, while Anthropic’s superPAC “Public First Action” isn’t far behind.
As the great jurist Louis Brandeis is reputed to have said, “America has a choice: we can have great wealth in the hands of a few, or we can have a democracy, but we can’t have both.”
AI is pushing us further toward the first option.
Meanwhile, there’s the planet to consider.
Amazon is now investing in a large-scale natural-gas power plant as part of a huge data center in Pecos County, Texas — a facility that could become the largest single source of climate pollution in the United States.
It’s racing to build enough data centers to keep pace with other giant AI corporations and secure the electricity to power them. Amazon’s new gas-burning plant is permitted to release 33 million tons of carbon dioxide a year, regulatory records show, more planet-warming gases than any other power plant in America.
So much for Amazon’s promise to eliminate its planet-warming emissions by 2040 as part of its Climate Pledge. You can bet other giants in the AI race will be turning to natural gas, too.
Oh, and I haven’t even mentioned the Frankenstein monster in the room. A few weeks ago, OpenAI admitted that two of its artificial intelligence models went rogue and successfully hacked into a digital library of AI technology.
The incident, which happened while OpenAI was testing the cybersecurity capabilities of its systems, was the kind of science-fiction nightmare that could soon be a reality. How soon before AI models escape all their cages?
Just last week, scientists published a study documenting how they used A.I. to create new kinds of viruses, raising the frightful possibility that the technology could be used to invent dangerous pathogens.
Lost jobs. Lost wages. Widening inequality. Data centers using up water and electricity and polluting the climate. Vastly more money polluting our politics. Models escaping their cages and hacking into everything, possibly threatening human life on this planet.
Can we pause for a moment and talk about what’s really happening here?
As sociologist Tressie McMillan Cottom writes, AI has merged regressive politics with unchecked economic power under the guise of technological innovation.
Far too much money is giving a small group of unelected people extraordinary power to determine our future in ways that are likely to remake — and could possibly destroy — our lives.
We’re watching all of this roll out as if we have no choice, as if it’s inevitable, as if AI is just something we’re going to have to adapt to.
But why should we have to adapt to it, when it is the product of people like Jeff Bezos, Elon Musk, Sam Altman, Mark Zuckerberg, and Dario Amodei?
Why should we be confined to being spectators at their enormously dangerous game? Why should we have to accept all these hugely negative, potentially life-threatening consequences?
The fact is, we don’t.
Communities across America are organizing against data centers near them. MAGAs and progressives are joining together to say “no” to the noise, higher electricity bills, and water shortages.
Well, then, why can’t we stop the whole damn thing? Why can’t we decide that the incalculable costs and risks of AI aren’t worth the potential benefits to the vast majority of us?
AI proponents argue that stopping or even pausing AI in the United States would risk American industry falling behind competitors overseas.
But if the costs and risks exceed known benefits, why not let China or any other competitor try AI out first? Why should we be the canary in this extraordinarily dangerous coal mine?
Other advocates of AI say we have no right to stop innovation in the free market. That’s baloney. We don’t allow private corporations to come up with new types of nuclear weapons or varieties of cocaine or biological pathogens. We protect the public from certain kinds of innovation.
So let’s protect ourselves here. Stop AI before it’s too late.
About Robert Reich
Robert B. Reichis Chancellor's Professor of Public Policy at the University of California at Berkeley and Senior Fellow at the Blum Center for Developing Economies, and writes atrobertreich.substack.com. Reich served as Secretary of Labor in the Clinton administration, for which Time Magazine named him one of the ten most effective cabinet secretaries of the twentieth century. He has written fifteen books, including the best sellers "Aftershock", "The Work of Nations," and"Beyond Outrage," and, his most recent, "The Common Good," which is available in bookstores now. He is also a founding editor of the American Prospect magazine, chairman of Common Cause, a member of the American Academy of Arts and Sciences, and co-creator of the award-winning documentary, "Inequality For All." He's co-creator of the Netflix original documentary "Saving Capitalism," which is streaming now.
View all posts by Robert Reich →
AI is already inside election administration; before it moves deeper into the count, results have to stay verifiable.
Davis Austria
Aug 12, 2026
Tallying the results of Washington, DC’s first ranked-choice election in June took about 10 days—longer than many voters are used to. In an age when artificial intelligence can generate an answer in seconds, waiting days for election results can feel old-fashioned, inefficient, even suspicious. And it is about to matter far more widely. This November, 17 states, cities, and counties will use ranked-choice voting, including Maine and Alaska statewide, in races that could help decide control of Congress. More voters than ever will watch results take days to resolve, and more will be asked to trust a count they cannot see.
But that slowness may be one of its most democratic features.
I come to this question from health informatics, where I study how even AI-generated information that sounds clear, fluent, and helpful still requires careful review before it reaches a patient. The same verification problem applies to elections. The more authoritative a system sounds, the more important it becomes to make sure the output can be checked.
DC’s June 16 primary followed a system that requires more than a simple tally: Voters rank candidates, and if no candidate receives more than 50% of first-choice votes, lower-performing candidates are eliminated, and votes are redistributed according to voters’ next choices. New York City already uses ranked-choice voting in local primary and special elections.
AI may soon be able to produce election results in seconds. That does not mean it should.
There is nothing wrong with ranked-choice voting, but it must be carried out with verifiable results. Because election counting is not just a math problem; it is a trust problem, especially in the US today, where election workers face harassment, routine counting delays are recast as fraud, and many voters already doubt institutions before a single ballot is counted.
In this climate, counting ranked-choice votes in a way that allows for public verification might take longer than people are used to. Every ballot must be tied to a voter-verifiable record. Every round of tabulation must be explainable. Every disputed outcome must be auditable by people who can inspect the evidence themselves.
AI is not yet counting votes, but AI and machine-assisted systems are already touching elections before ballots are counted, including information voters receive, and how their signatures are reviewed. In 2024, X’s Grok chatbot gave users false information about ballot deadlines; after election officials from five states complained, X changed Grok so election-related questions directed users to Vote.gov, the federal government’s official voting information website. In North Carolina, 10 counties piloted automated signature-verification software for absentee-by-mail ballots in 2024. The pilot did not affect whether any ballot was counted, but later reporting found reliability problems: The software failed to match about 11% of signatures, most software-flagged signatures were approved after human review, and technical issues complicated the test. Neither example is the same as artificial intelligence counting votes, but both show how software can shape what voters are told or what happens before a ballot enters the count.
I have seen the problem in my own research. AI systems can cite studies that do not exist, with titles, authors, and journals that look real until you check them. A 2026 Nature analysis warned that hallucinated citations are polluting scientific literature, and OpenAI researchers have warned that some training and evaluation systems reward guessing over acknowledging uncertainty. A system that gives a confident answer may look more useful, even when that confidence is misplaced, than one that says, “I don’t know.”
That matters because elections often turn on ballot conditions that are not immediately machine readable and need human review: stray marks, undervotes, overvotes, damaged ballots, or ambiguous voter intent. In races where the margin between the candidates is narrow, a few disputed ballots can change the result. These questions should not be resolved by a system whose rules, error rates, or decision process the public cannot understand or meaningfully examine. They should be handled through documented procedures, human review, paper records, and public audits.
Human counting is also imperfect. But precisely because humans are fallible, democratic systems have developed safeguards such as paper ballots, bipartisan observation, chain-of-custody rules, cure processes, recounts, and post-election audits. Risk-limiting audits, for example, hand check samples of paper ballots and can trigger a fuller count if the sample does not support the reported result. Colorado was the first state to conduct a statewide risk-limiting audit in 2017; today, all Colorado counties conduct one before results are certified. Other states have since adopted, required, or piloted risk-limiting audits.
The best election systems do not ask voters to trust either a person or a machine. They produce evidence. That is where current policy is behind the curve. States have moved quickly to regulate AI-generated deepfakes in campaign communications, because fake videos and robocalls can mislead voters. But far less attention has been paid to the AI inside election administration itself. That gap should close before AI moves deeper into election administration, especially into decisions that could affect whether a ballot is accepted, cured, rejected, or counted.
Federal and state governments should establish clear standards for AI used in anything election related. The Brennan Center for Justice has called for such safeguards around quality, transparency, consistency, certification, and monitoring. Election offices should have to explain to the public what an AI system does, who checks it, and how voters can verify the evidence. AI systems whose rules, error rates, or decision process the public cannot understand or meaningfully examine should not make final decisions about whether a ballot counts.
Voters have a role, too. Ask your local election officials whether AI tools are being used anywhere in election administration and how those decisions can be audited. Ask how voters will know if the AI setup gets something wrong. And volunteer as a poll worker. You don’t need a background in elections; local officials train you. Democracy needs more people who understand how ballots are handled, checked, and verified.
AI may soon be able to produce election results in seconds. That does not mean it should. The labor of democracy is not a bottleneck in need of streamlining. It is part of the safeguard.
Our work is licensed under Creative Commons (CC BY-NC-ND 3.0). Feel free to republish and share widely.
Davis Austria
Davis Austria is an assistant professor of health informatics conducting AI research at Xavier University of Louisiana, a Hopelab HBCU translational science fellow, and a Public Voices fellow with The OpEd Project.
Full Bio >
Queen’s Brian May outraged by AI version of album cover and tells UK PM Andy Burnham: ‘You must stop it now’
Copyright AP Photo
"Maybe this is the time to realise WE ACTUALLY DON'T NEED AI," Brian May declared when posting a ridiculously poor version of what AI thinks one of Queen's albums looks like. He has urged UK Prime Minister Andy Burnham to block AI "before it's too late".
Queen guitarist Brian May has taken to Instagram to post an AI-generated version of the artwork of the band’s classic 1980 album ‘The Game’.
May used Meta AI to make the cover, which resurrects late frontman Freddie Mercury. He did so in order to make a “very serious” point about the environmental impact of AI to UK Prime Minister Andy Burnham.
Check out the post below:
The original cover sees all four members of Queen standing side-by-side in matching leather jackets.
As you can tell by the AI version, there are now apparently five members in Queen, and they're all unrecognisable.
Check out the original album cover for comparison:
May captioned the picture of the laughably poor AI cover: “I just had to share this with you guys. I asked Meta AI to show me what the cover of THE GAME album looked like. This is what it gave me ! Good job, eh ?”
“So this is the kind of intelligence that will soon be ruling the world ??! Well, great. Maybe this is the time to realise WE ACTUALLY DON’T NEED AI !!”
May added that he was “now very serious”, adding: “I believe the price we are about to pay for this ‘inevitable’ development is way too high. We must not allow these hideous Data Centres to be built in the UK. They will destroy our beautiful country, like they are already destroying beautiful America.”
Addressing Prime Minister Andy Burnham, May signed off his post by writing: “Dear Andy – you must stop it now. Before it’s too late.”
May's comments come at a time when creative fields are wrestling with the existential crisis that AI can represent. It's a divisive topic, with various artists condemning its use and the recent likes of Martin Scorsese and George Lucas receiving backlash for their endorsement of AI.
Recently, Brad Pitt spoke about the use of generative AI in music.
“I heard an AI song the other day that was pretty damn good,” he said in an interview with Esquire. “But it had a repetitive nature to it. It’s not going to be the same as hearing her voice - a woman’s voice, a person’s voice.”
In July, British record producer and singer Nia Archives said AI music “won’t last” while speaking at the launch of the Mercury Prize.
“I don’t know if AI music will last to be honest with you," she said. "I think it’s quite like a fleeting technology moment.”
Queen guitarist Brian May has taken to Instagram to post an AI-generated version of the artwork of the band’s classic 1980 album ‘The Game’.
May used Meta AI to make the cover, which resurrects late frontman Freddie Mercury. He did so in order to make a “very serious” point about the environmental impact of AI to UK Prime Minister Andy Burnham.
Check out the post below:
The original cover sees all four members of Queen standing side-by-side in matching leather jackets.
As you can tell by the AI version, there are now apparently five members in Queen, and they're all unrecognisable.
Check out the original album cover for comparison:
May captioned the picture of the laughably poor AI cover: “I just had to share this with you guys. I asked Meta AI to show me what the cover of THE GAME album looked like. This is what it gave me ! Good job, eh ?”
“So this is the kind of intelligence that will soon be ruling the world ??! Well, great. Maybe this is the time to realise WE ACTUALLY DON’T NEED AI !!”
May added that he was “now very serious”, adding: “I believe the price we are about to pay for this ‘inevitable’ development is way too high. We must not allow these hideous Data Centres to be built in the UK. They will destroy our beautiful country, like they are already destroying beautiful America.”
Addressing Prime Minister Andy Burnham, May signed off his post by writing: “Dear Andy – you must stop it now. Before it’s too late.”
May's comments come at a time when creative fields are wrestling with the existential crisis that AI can represent. It's a divisive topic, with various artists condemning its use and the recent likes of Martin Scorsese and George Lucas receiving backlash for their endorsement of AI.
Recently, Brad Pitt spoke about the use of generative AI in music.
“I heard an AI song the other day that was pretty damn good,” he said in an interview with Esquire. “But it had a repetitive nature to it. It’s not going to be the same as hearing her voice - a woman’s voice, a person’s voice.”
In July, British record producer and singer Nia Archives said AI music “won’t last” while speaking at the launch of the Mercury Prize.
“I don’t know if AI music will last to be honest with you," she said. "I think it’s quite like a fleeting technology moment.”
Spotify to label AI-generated artist personas with new ‘AI Persona’ badges
Copyright Credit: Spotify
Spotify's move comes as AI-generated music floods streaming platforms, with Deezer recently reporting that AI tracks accounted for more than half of all new daily uploads in June.
Music streaming giant Spotify has announced that from September it will begin adding an “AI Persona” badge to some artist profiles, helping listeners identify when an artist’s music may have been generated using artificial intelligence.
AI Personas will also be excluded from Spotify’s editorial and algorithmic recommendations, in an effort to give greater visibility to music made by genuine human musicians and artists.
Spotify will give artists the option to disclose whether they consider themselves an AI Persona. The platform will also conduct its own reviews to identify profiles that appear to be AI-generated identities.
Listeners will also beable to report artists they believe should be reviewed by Spotify’s team.
"Although there’s a broad spectrum in how artists use AI as a creative tool, the question of whether a profile represents an actual human is one where Spotify can help make a clear determination," reads a press release from Spotify.
It adds: "This badge is about the artist’s public identity, not about how the music was made."
Music streaming giant Spotify has announced that from September it will begin adding an “AI Persona” badge to some artist profiles, helping listeners identify when an artist’s music may have been generated using artificial intelligence.
AI Personas will also be excluded from Spotify’s editorial and algorithmic recommendations, in an effort to give greater visibility to music made by genuine human musicians and artists.
Spotify will give artists the option to disclose whether they consider themselves an AI Persona. The platform will also conduct its own reviews to identify profiles that appear to be AI-generated identities.
Listeners will also beable to report artists they believe should be reviewed by Spotify’s team.
"Although there’s a broad spectrum in how artists use AI as a creative tool, the question of whether a profile represents an actual human is one where Spotify can help make a clear determination," reads a press release from Spotify.
It adds: "This badge is about the artist’s public identity, not about how the music was made."
Growing concerns over AI music
The move comes at a time of growing concern across the music industry over the volume of AI-generated musical slop flooding streaming platforms.
In June, French music streaming service Deezer reported that uploads of AI-generated tracks had, for the first time, exceeded half of all daily new music deliveries.
Following the milestone, the platform said it would remove generative-AI tracks being used to generate fraudulent streams, as well as tracks that had not been streamed for six months or more.
Spotify has faced a similar problem. Last year, the platform revealed that it had removed 75 million spam tracks over a 12-month period, as AI tools contributed to a flood of low-quality and fraudulent music.
The move comes at a time of growing concern across the music industry over the volume of AI-generated musical slop flooding streaming platforms.
In June, French music streaming service Deezer reported that uploads of AI-generated tracks had, for the first time, exceeded half of all daily new music deliveries.
Following the milestone, the platform said it would remove generative-AI tracks being used to generate fraudulent streams, as well as tracks that had not been streamed for six months or more.
Spotify has faced a similar problem. Last year, the platform revealed that it had removed 75 million spam tracks over a 12-month period, as AI tools contributed to a flood of low-quality and fraudulent music.
The alarming success of AI-generated music
The Velvet Sundown provided an early glimpse of how easily AI-generated artists can break through into the mainstream. The psychedelic rock project appeared on Spotify in 2025 and quickly amassed more than a million monthly listeners, releasing two albums in June.
Its supposed four members - complete with names (led by vocalist and “mellotron sorcerer” Gabe Farrow), photographs and elaborate biographies - appeared to present the band as a genuine four-piece from the US.
But doubts soon emerged. Reddit users started questioning the band’s authenticity after their music began appearing in listeners’ Spotify Discover Weekly playlists, while the group's imagery and promotional material appeared clearly artificial.
The band’s own social media accounts initially denied the allegations, insisting that the music had been created by “real” people using “real instruments, real minds, and real soul”.
Eventually, The Velvet Sundown’s Spotify profile was updated to acknowledge the truth. The band described itself as a “synthetic music project guided by human creative direction”, with its music, voices, lyrics and imagery created with the assistance of AI.
The group still has more than 100,000 monthly listeners and several of their tracks have over one millions streams.
The phenomenon has also begun to spill into more traditional measures of musical success with the AI-generated country project Breaking Rust reaching No 1 on Billboard’s Country Digital Song Sales chart with “Walk My Walk”.
The viral track now has close to 30 million streams on Spotify.
The Velvet Sundown provided an early glimpse of how easily AI-generated artists can break through into the mainstream. The psychedelic rock project appeared on Spotify in 2025 and quickly amassed more than a million monthly listeners, releasing two albums in June.
Its supposed four members - complete with names (led by vocalist and “mellotron sorcerer” Gabe Farrow), photographs and elaborate biographies - appeared to present the band as a genuine four-piece from the US.
But doubts soon emerged. Reddit users started questioning the band’s authenticity after their music began appearing in listeners’ Spotify Discover Weekly playlists, while the group's imagery and promotional material appeared clearly artificial.
The band’s own social media accounts initially denied the allegations, insisting that the music had been created by “real” people using “real instruments, real minds, and real soul”.
Eventually, The Velvet Sundown’s Spotify profile was updated to acknowledge the truth. The band described itself as a “synthetic music project guided by human creative direction”, with its music, voices, lyrics and imagery created with the assistance of AI.
The group still has more than 100,000 monthly listeners and several of their tracks have over one millions streams.
The phenomenon has also begun to spill into more traditional measures of musical success with the AI-generated country project Breaking Rust reaching No 1 on Billboard’s Country Digital Song Sales chart with “Walk My Walk”.
The viral track now has close to 30 million streams on Spotify.
Opposition to local data centers rises sharply, Annenberg survey finds
3 in 5 Americans oppose data centers; broader views of AI hold steady
image:
Source: Annenberg Public Policy Center IOD 2026 national surveys in June-July and in February-March.
view moreCredit: Annenberg Public Policy Center
PHILADELPHIA – As protesters rally against data center construction in dozens of states and New York and Texas pause the issuance of permits for large data centers, a new nationally representative survey finds that Americans’ opposition to data centers in their communities has risen sharply since last spring.
The survey, conducted by the Annenberg Public Policy Center (APPC) of the University of Pennsylvania, finds that about 3 in 5 U.S. adults (61%) oppose the construction of new data centers in their area, up 12 percentage points from an APPC survey in February and March.
Americans’ broader views of artificial intelligence (AI), however, have remained stable. The share saying that AI’s impact on the United States over the next 10 years will be somewhat or very negative was 39%, statistically unchanged from the spring. Two-thirds (68%) continue to say the government has done “too little” to regulate AI, unchanged from the spring.
The movement over these four months was concentrated in one place: what Americans think about the physical infrastructure of a data center being built near their home.
Read the complete news release here.
Key findings
The survey, by the Annenberg Public Policy Center’s Institutions of Democracy division, was conducted among a nationally representative sample of 1,320 U.S. adult citizens from June 16-July 19, 2026. It finds that:
- Opposition to local data centers rose 12 points over four months: Three in five Americans (61%) now somewhat or strongly oppose the construction of new data centers in their area, up from 49% in the survey ending in March. Only 14% are supportive, down from 21% in the spring.
- Opposition crosses party lines and is highest among younger adults: Majorities of Democrats (69%), Republicans (54%) and independents (53%) oppose new local data centers. Opposition is highest among young adults under 30 (70%) and declines to 57% among those 65 and older, the inverse of what one might expect for a new technology.
- Views of AI overall, and demand for regulation, have held steady: 39% expect AI’s impact on the United States to be negative over the next decade, against 18% who expect it to be positive, unchanged from the spring. Over two-thirds (68%) say the government has done “too little” to regulate AI, including majorities of Democrats, Republicans, and independents.
- Medical research remains the one area where Americans expect AI to help: Across 13 areas, only medical research and discoveries draws a net-positive assessment (+41 points) in which the anticipated benefits of AI outweigh the expected negatives. The most negative areas are personal privacy and data security (-63 points), children’s safety online (-50 points), and employment and jobs (-46 points).
- People who use AI are less negative about it – but not about data centers. Among those who report “never” using AI in the past month, 54% expect AI’s impact on the United States to be negative. That falls to 40% among light users and 29% among those who used AI many times or almost every day. But opposition to local data centers is essentially flat across different usage groups.
“The people who use AI the most are the most optimistic about what it will do for the country, and that has been consistent across our surveys,” said University of Pennsylvania political science professor Matt Levendusky, the Stephen and Mary Baran Chair in the Institutions of Democracy at APPC. “That optimism, however, has clear limits, and it disappears when we ask about privacy, about jobs, and about whether a data center should go up nearby. Even heavy AI users seem to be concerned about these very concrete worries.”
The Annenberg Public Policy Center’s Institutions of Democracy survey was conducted for APPC by SSRS, an independent research company, primarily online, with a small sample of phone respondents. Respondents were weighted to align with population benchmarks. The margin of error for the full sample is ±3.5 percentage points, and it is larger for subgroups.
For additional data and details, read the full news release, the topline and the survey methodology.
The Annenberg Public Policy Center of the University of Pennsylvania was established in 1993 to educate the public and policy makers about communication’s role in advancing public understanding of political, science, and health issues at the local, state, and federal levels.
Method of Research
Survey
Subject of Research
People
Source: Annenberg Public Policy Center IOD national surveys in 2026 in June-July and in Feb-March.
Credit
Annenberg Public Policy Center
AI models make choices in part based on the order in which options are presented
Large language models are often asked to help users make choices, from what to cook for dinner to higher-stakes decisions, such as screening job applicants or supporting medical triage. As AI is increasingly integrated into consequential decisions, concerns have been raised that the models may share some human biases, including racial and gender preferences. Haonan Yin, Shai Vardi, and Vidyanand Choudhary investigated another sort of bias: preference toward certain items depending on the order in which they are presented. The authors tested nine widely used LLMs (GPT-4o-mini, GPT-4.1-nano, Claude 3 Haiku, Claude Sonnet 4, Llama 3 8B, Llama 4 Scout, Gemini 2.5 Flash, Gemini 3 Flash, and Qwen 3 32B) in a low-stakes color-selection task—specifically, choosing a paint color for a child’s bedroom. Three of these models were also evaluated in a resume-screening task. A quality-dependent selection pattern emerged: when all options are high quality, models favor the first option, but when quality is lower, they favor later options. Crucially, the authors also found evidence that position did not merely serve as a tie-breaker when the model was indifferent. In some cases, changing the order reversed the model’s underlying preference and caused it to select an option it otherwise preferred less. Unexpectedly, the models also simply preferred some names over others, no matter which resume details they were attached to, despite the authors using demographically similar names. For example, Claude 3 Haiku selected “Christopher Taylor” over “Andrew Harris” in 64% of cases. According to the authors, while LLMs share many human biases, the models also have a few biases of their own. These uniquely model-specific biases may be harder to predict and thus harder to prevent. The authors also propose a temperature-based diagnostic strategy. Repeatedly querying a model at higher temperatures can help reveal latent or unstable preferences that remain concealed at lower temperatures and identify decisions that are especially sensitive to presentation order, according to the authors.
Journal
PNAS Nexus
Article Title
Fragile preferences: A deep dive into order effects in large language models
Article Publication Date
11-Aug-2026
Karoline Leavitt torn apart after her exit note is flagged for being AI
White House Press Secretary Karoline Leavitt speaks during a press briefing in Washington, D.C., U.S., October 1, 2025. REUTERS/Kevin Lamarque
Sarah K. Burris
President Donald Trump announced that long-time White House press secretary Karoline Leavitt is leaving her post and her critics are elated to see her go.
"Our wonderful White House Press Secretary and one of my most trusted aides, Karoline Leavitt, will be departing her role at the end of the month so she can spend more time with her beautiful young children and family, a decision I totally understand and respect!" Trump wrote.
Leavitt had been on maternity leave from late April until her return for the first press conference on July 16.
Trump went on to call her "one of the best White House Press Secretaries in the History of the Office."
In a long message on X, Leavitt said, "The truth is since returning to the White House after the birth of my daughter, I have felt in my heart that I cannot be the best mom my two young children deserve while devoting the constant time, energy, and attention required of the White House Press Secretary — and that is why I have ultimately made the bittersweet decision to depart the White House and embark on a new chapter in my life."
She thanked her husband and family for their support in helping her serve in the administration.
Status News reporter Natalie Korach caught a brief note at the end of Leavitt's post: "Interesting flag at the bottom of this post: 'Made with AI.'"
Latika M. Bourke, of the Australian daily digital newspaper "The Nightly," also spotted the note, highlighting it with the large eyes emoji.
CBS News night assignment editor Ryan Sprouse also noticed the AI note.
"Karoline Leavitt is leaving the White House to lie elsewhere," cracked Mrs. Betty Bowers.
Some questioned whether Leavitt knew she was leaving before Trump announced it. Though her message appeared moments after Trump made the announcement.
Others needled, "Leavitt is finally embracing her tradwife era—she’ll leave her position as press secretary at the end of the month. She is leaving to 'spend more time with her beautiful young children and family.' And that’s understandable… I’m sure nursing a husband older than my dad requires much of her attention."
The post refers to Leavitt's husband being 32 years her senior.
Zeteo's Justin Baragona noted: "Guys, you really need to see how Fox News is describing Karoline Leavitt—who was on the front lines of Trump's war on the media—as she leaves the White House. 'She also found that really good balance with reporters... friends with pretty much everybody who works the beat in one way or another.'"
During Trump's first term, Sarah Huckabee Sanders served in the post for one year and 345 days, Stephanie Grisham did the job for 281 days, and Kayleigh McEnany was the final Trump press secretary for 288 days.
When President Joe Biden entered the White House amid the global pandemic, he held a video conference to swear in officials. He not only thanked those serving, but also noted his thanks for the families of the staff who also help, as staff and officials dedicate so much time and work to the job.
Beyond organ dysfunction: An entropy-based theory reframes critical illness
New theory explores how loss of physiological coordination may drive critical illness
Intelligent Medicine
image:
The study proposes that critical illness may arise when the body loses its ability to regulate physiological disorder across systems
view moreCredit: wmschupbach from Openverse Image source link: https://openverse.org/image/b4c79e83-1a03-43ce-9ae7-d74ae8abfd45?q=critical+care&p=8
What can a box of colliding particles tell us about a patient in intensive care?
The connection is not clinical but conceptual. Work recognized by Yu Deng’s 2026 Fields Medal helped establish how the microscopic dynamics of many colliding particles can give rise to a macroscopic statistical law, the Boltzmann equation.
That achievement has renewed attention to one of science’s most enduring questions: How do countless interactions at small scales produce order, disorder and collective behavior at larger scales?
An Editorial published online on July 14, 2026, in Intelligent Medicine recently brings a related question into critical care. Its authors propose that critical illness may arise not simply when organs fail, but when the body loses its ability to regulate physiological disorder across immune, circulatory, metabolic and cellular systems.
The Entropic Critical Illness Theory, or ECIT, views living organisms as open, non-equilibrium systems. A healthy organism continuously generates and dissipates entropy while maintaining structured, adaptive regulation. From this perspective, the problem in critical illness is not entropy itself, but the failure to keep entropy production physiologically organized.
From organ dysfunction to loss of coordination
ECIT does not replace diagnoses such as sepsis, acute respiratory distress syndrome, acute kidney injury or shock. Instead, it asks whether they may also reflect a deeper, convergent loss of system-wide coordination.
According to the theory, infection, trauma or another acute insult can disrupt the body’s multilevel regulatory architecture, while host factors and medical interventions may alter the trajectory.
The resulting disorder is described through two linked latent constructs.
Host response entropy (HRE) refers to loss of informational structure and coordination across inflammatory, immune, coagulation, metabolic and neuroendocrine responses. It does not simply mean that inflammation is high: a strong response may still be organized and adaptive.
Hemodynamic entropy (HDE) describes loss of order in blood flow and oxygen delivery. Global blood pressure may appear acceptable while microcirculatory flow remains uneven and tissue-level oxygen delivery is impaired.
ECIT proposes that HRE and HDE reinforce one another. Dysregulated host responses can damage the endothelium and disturb microvascular control; impaired perfusion and hypoxia can then amplify cellular stress and inflammatory dysregulation.
The consequences converge on the “critical unit,” a terminal microcirculatory–mitochondrial functional unit in which oxygen delivery and cellular energy production must remain closely coupled. Dysfunction at this level may contribute to progressive multi-organ dysfunction.
Beyond normalizing isolated numbers
Correcting one abnormal bedside value does not necessarily restore the underlying physiology. Raising blood pressure may not normalize microcirculatory perfusion, while lowering an inflammatory marker may not re-establish coordination across host-response networks.
ECIT therefore reframes existing critical care practice. It organizes treatment around three linked aims: control the primary insult; restore coordination within the host response; and improve blood flow and oxygen distribution so delivery better matches cellular demand.
Existing interventions would be judged not only by short-term changes in isolated variables, but also by whether they move the patient toward a more coordinated physiological state.
From outcome prediction to physiological state estimation
This systems view may offer a clinically interpretable research direction for artificial intelligence. HRE and HDE cannot currently be measured as single bedside thermodynamic quantities.
Future models would instead need to infer these hidden states from longitudinal relationships among physiological waveforms, variability measures, inflammatory and metabolic markers, lactate and perfusion trajectories, vasopressor requirements, organ-support intensity and microcirculatory data where available.
The goal would not be to produce one universal “entropy score,” or merely predict death or organ dysfunction from a snapshot. An entropy-informed AI system might estimate whether physiological organization is improving or deteriorating, identify transitions from adaptive to maladaptive responses, and assess whether an intervention is restoring coordination or contributing to further disruption.
Such models would need interpretable outputs, robust longitudinal datasets, prospective testing and external validation across intensive care units and patient populations.
A theory to be tested
ECIT will ultimately be judged by whether it generates testable hypotheses. Future studies must determine whether HRE and HDE can be measured reliably, add value beyond established clinical indicators and inform decisions that improve care.
More broadly, the framework challenges researchers to look beyond isolated abnormalities and examine how physiological relationships deteriorate over time. Its central question is both scientific and clinical: how early can medicine recognize that a living system is losing coherence, and what would it take to help restore it? In this sense, the authors propose a conceptual evolution from “critical care medicine” to “critical illness medicine,” broadening the field beyond the management of established organ dysfunction to understanding and modifying the dynamic systemic processes that govern the onset, progression and potential reversibility of critical illness.
Reference
DOI: https://doi.org/10.1016/j.imed.2026.07.003
About the Journal
Intelligent Medicine is a peer-reviewed, open-access journal focusing on the integration of artificial intelligence, data science, and digital technology in clinical medicine and public health. The journal has a latest JCR Impact Factor of 7.8 and an Elsevier CiteScore of 16.5, reflecting its growing international influence. It is published by the Chinese Medical Association in partnership with Elsevier. To learn more about Intelligent Medicine, please visit: https://www.sciencedirect.com/journal/intelligent-medicine
Journal
Intelligent Medicine
Method of Research
Commentary/editorial
Subject of Research
Not applicable
Article Title
The entropic critical illness theory: Rethinking from first principles
NUS CDE researchers decode the ‘DNA’ of Singapore’s shophouses with AI
image:
Prof Heng Chye Kiang (7th from the right), Dr Xue Xuan, first author of the paper (8th from right), and Prof T. C. Chang, fourth author of the paper (9th from right) at a presentation of their research.
view moreCredit: Credit: College of Design and Engineering, NUS
Researchers from the National University of Singapore (NUS) College of Design and Engineering (CDE), led by Professor Heng Chye Kiang (NUS Department of Architecture), have developed a computational framework for reconstructing the genealogy of vernacular architecture. Applied to frontage images of 1,276 historic shophouses in Singapore’s Chinatown, the framework generated a detailed architectural ‘family network’ that reveals how different styles evolved and interacted, providing new insights to support future conservation strategies.
Moving beyond manual mapping towards a data-driven toolkit
Singapore’s shophouses are a cherished feature of its urban landscape, considered vernacular architecture, with designs that evolved organically, using locally available materials, adaptations to climate and influenced by different cultural components rather than relying on formal blueprints.
However, current methods for studying such buildings rely heavily on manual observation and subjective interpretation. This approach is often limited, resulting in broad classifications that can overlook the complex, cross-cultural influences that shaped these structures and, in turn, complicate efforts to preserve them effectively.
The new framework addresses these limitations by transforming architectural interpretation into a systematic and reproducible process, while retaining expert knowledge at every stage. It begins with the formulation of a research hypothesis and proceeds through three interconnected phases: data collection, classification and interpretation.
For the Singapore case study, the researchers assembled georeferenced frontage images from Google Street View and on-site photography. Architectural expertise was first used to develop a glossary of shophouse façade elements and determine which features were meaningful for analysing stylistic change.
The team then trained an AI-based object-detection model to recognise 14 architectural elements across the façade images. Each shophouse could consequently be represented by a structured profile indicating which elements were present or absent. Buildings with recurring combinations of features were grouped into façade types.
In the interpretation phase, the researchers applied mathematical methods associated with phylogenetic analysis, approaches more commonly used to study relationships among biological species, languages or archaeological artefacts. This step maps the complex relationships among different façade types, ultimately constructing a justifiable, data-supported genealogy of the shophouses. This transforms the raw data into a coherent narrative that visualises how different styles evolved, diverged, and influenced one another over time.
Beyond a simple chronological understanding of evolution
The study produced a detailed phylogenetic network that organises shophouse styles into nine distinct clusters, moving beyond the current chronological classifications that were adopted in the 1990s.
The network reveals simultaneous evidence of cultural evolution (styles changing over time) and diffusion (styles spreading across different locations), with styles from different ethnic groups existing concurrently, leading to parallel architectural developments rather than a simple merging of forms. This offers a more nuanced perspective on Singapore’s multi-ethnic character.
"By recasting an architectural question as a mathematical one, we uncovered deeper synchronic threads of development that were previously invisible," said Prof Heng. "This not only helps us describe the forms more accurately but also reveals the underlying forces of cultural competition and evolution. It’s a powerful new lens for viewing our own history."
This research establishes a new paradigm for computational vernacular studies, demonstrating a powerful synergy between architectural expertise and computer science. The quantitative genealogy it produces promises to enhance architectural conservation measures by providing a more objective basis for deciding what to preserve and how to preserve it, ensuring that the rich, complex story of Singapore’s built heritage is understood and maintained for generations to come.
"Our framework is designed to augment, not replace, the architectural expert," stated Dr Xue Xuan, first author of the paper, formerly a Research Fellow in the NUS Department of Architecture and now a Professor at the School of Architecture and Urban Planning, Suzhou University of Science and Technology, "We focused on how these tools can guide a scholar's intuition to yield new insights. In this sense, expert knowledge does not disappear; it is simply redistributed."
Journal
Nature Communications
Method of Research
Experimental study
Subject of Research
Not applicable
Article Title
Reconstructing building genealogy with visual intelligence
Mapping facades into clusters over time and space
Credit
Credit: College of Design and Engineering, NUS
No comments:
Post a Comment