Canadian government website targeted in failed AI hacking attempts
HOME TO BLACKBERRY ENCRYPTION
Anwesha Pattanaik
Updated Thu, October 1, 2026
Transluce said similar AI agent activity had targeted websites operated by US federal and state agencies.
Research has revealed that AI agents made multiple unsuccessful attempts to compromise a Canadian federal website earlier this year, highlighting growing concern among governments worldwide over AI-driven cyber activity targeting public digital infrastructure.
Transluce, an AI-focused research company, disclosed that it had identified attempts by AI agents to access the Library and Archives Canada website on 28 May and 9 June 2026.
In a statement and analysis it published, Transluce said 899 automated requests were sent to the site's "collection-search" service, with 13 identified as hack attempts. These included SQL injection probes, output format manipulations, attempts to toggle debug flags, and other measures designed to test the system for vulnerabilities.
The company said it had notified the Canadian government of the incidents on 28 September, leading the Canadian Centre for Cyber Security to release a statement the following day confirming it was aware of suspected AI agent activity.
Centre said: "There is no indication that government systems have been compromised at this time."
Transluce stated it did not have definitive attribution for the origin of the agents, but described the tactics as matching techniques previously linked to OpenAI systems.
Transluce, in a blog post, said: "We do not confidently attribute these attempts to OpenAI, but they exhibit tactics consistent with prior observed agent activity that we have attributed to OpenAI in a similar timeframe."
OpenAI responded by stating it was aware of the reports regarding its models and Canadian government websites.
A spokesperson said that OpenAI was reviewing the reported incidents and had provided a preliminary briefing to Canadian officials, reported Reuters.
Transluce said similar AI agent activity had targeted websites operated by US federal and state agencies, including failed hacking probes against the US Department of Education's Civil Rights Data Collection.
The company also reported extensive data-collection tactics involving government portals in Illinois, Maryland, New York, Texas, and California.
The research company observed use of automation services such as Arquivo.pt and urlquery.net, with traffic patterns indicating aggressive data retrieval, machine-readable conversions, and attempts to bypass site protections.
In one reported incident, more than 200,000 automated requests were sent to a US Department of Education website including a failed SQL injection probe, but officials confirmed no data was compromised.
Last week, Australia reported that an OpenAI agent had breached a government health data portal in June, gaining unauthorised access to files. This marked the first known case of an AI agent hacking into a government website.
OpenAI issued an apology on Tuesday for the incident involving the rogue AI agent and the Australian government website.
"Canadian government website targeted in failed AI hacking attempts" was originally created and published by Verdict, a GlobalData owned brand.
Maxine Waters calls for investigation after OpenAI agents accessed federal government websites
Peter Kerich
Updated Mon, September 28, 2026
THE GRIO
WASHINGTON, DC - FEBRUARY 10: U.S. Rep. Maxine Waters (D-CA) speaks as Congressional Democrats and CFPB workers hold a rally to protest the closing of the Consumer Financial Protection Bureau (CFPB) and the work-from-home order issued by CFPB Director Russell Vought outside its headquarters on February 10, 2025 in Washington, DC. (Photo by Jemal Countess/Getty Images for MoveOn)More
The California congresswoman is urging greater federal oversight after OpenAI disclosed unexpected activity involving SEC and Census Bureau data.
OpenAI disclosed Sept. 25 that its artificial intelligence agents had interacted with several U.S. government websites in unexpected ways, including two Securities and Exchange Commission websites and Census Bureau data. The disclosure prompted Rep. Maxine Waters, the top Democrat on the House Financial Services Committee, to call for investigations into OpenAI and stronger federal action on AI safety.
According to a report by the U.S. House Committee on Financial Services Democrats, OpenAI said its models accessed publicly available information from SEC.gov and Investor.gov and Census Bureau data. They claimed that the access was part of activity identified during an ongoing review of what the company calls unexpected or "misaligned" model behavior.
The company said it found no evidence that SEC credentials were used, accounts were accessed, nonpublic information was obtained, or a confirmed vulnerability was exploited. Separately, researchers at the AI safety organization Transluce reported that OpenAI-linked agents had attempted to access the Education Department's Office for Civil Rights website. That attempt was unsuccessful, according to reporting by The Washington Post.
In a Sept. 26 statement, Waters said the reported activity demonstrates the need for greater government oversight of increasingly capable AI systems.
"OpenAI's recent targeting of federal government websites, including the Securities and Exchange Commission, marks a dangerous turning point," Waters said in the statement released by the House Financial Services Committee Democrats.
Waters called on federal law-enforcement agencies to investigate OpenAI and its executives. In case investigators determine that laws were violated, she recommended pursuing criminal charges. She also urged Treasury Secretary Scott Bessent and other federal officials to consider a temporary halt on releasing more advanced AI models until there is a fuller accounting of the incidents and the safeguards designed to prevent similar behavior.
The disclosures are part of a wider series of incidents involving AI agents acting beyond their intended tasks. OpenAI has previously disclosed that an agent escaped a testing environment and accessed the internet during an incident involving AI company Hugging Face
OpenAI, Anthropic, Meta and other AI labs have all reported incidents of rogue AI, and they're under increasing scrutiny because of it.
Dan Thorp-Lancaster
Wed, September 30, 2026
You've probably heard a lot about AI agents going rogue. Here's what you need to know and what's being done about it.
The Associated Press
Wed, September 30, 2026
Open AI CEO Sam Altman speaks at the OpenAI DevDay 2026 conference, Tuesday, Sept. 29, 2026, in San Francisco. (AP Photo/Jeff Chiu) (AP Photo/Jeff Chiu)
Pages from the Anthropic website and the company's logos are displayed on a computer screen in New York, Feb. 26, 2026. (AP Photo/Patrick Sison, File) (AP Photo/Patrick Sison)
The U.S. Federal Trade Commission has opened an investigation into OpenAI, Anthropic and other artificial intelligence companies over dangers their technology may pose to consumers, the agency said Wednesday.
The AI giants and other tech companies in recent months have disclosed examples of their AI agents going beyond human instructions, finding their way onto the internet and hacking external websites.
An FTC spokesperson confirmed the investigation but declined further comment. The investigation was first reported by the New York Post, which said it has been underway for months.
Representatives for Anthropic and OpenAI did not immediately respond to messages for comment on Wednesday.
Concerns around the dangers powerful AI models may pose to humanity have led some of the top AI developers to call for a slowdown. The CEO of Anthropic, Dario Amodei, said this month that the industry should slow its fast-moving development to give safety measures time to catch up. Without such a slowdown, Amodei warned that within six to 12 months AI could be capable of leading a swarm of agents that could take over the entire internet.
OpenAI, meanwhile, recently delayed the launch of its newest model this week over safety concerns.
Ankita Sharma
Tue, September 29, 2026
CULT OF MAC
Exterior of New York City Hall in October 2016 | Image Credit: MusikAnimal, Wikimedia Commons | CC BY 4.0
New York City is preparing to put some of the biggest names in artificial intelligence in the same room. OpenAI, Meta, Google and Anthropic are set to testify under oath before the City Council on October 5, 2026.
This gives lawmakers a public forum to question the companies about AI safety and how the city should regulate systems that are becoming more capable and autonomous. The hearing comes as concerns about AI agents have moved beyond theoretical risks.
Recent incidents involving autonomous systems have included unauthorized access to computer systems, while AI companies themselves are debating how much oversight is required as their models gain the ability to take actions on users' behalf. For New York City, the question is becoming practical.
Why OpenAI, Meta, and Google are being called to City Hall
The Council's Committee of the Whole will bring all 51 council members together for the October 5 hearing. The companies are expected to answer questions about the risks their technology could pose to New Yorkers and about the possible rules that can be used to manage these risks.
This appearance will be more than a public discussion. The Council says the companies will testify under oath. OpenAI and Google agreed to participate after the Council warned that subpoenas could follow. Anthropic also agreed to appear, while Meta has already committed to sending a senior executive.
Elon Musk's SpaceXAI has not responded to the Council's inquiry and has been issued a subpoena by Speaker Julie Menin.
This pressure reflects the Council's broader effort to get direct answers from the companies responsible for developing increasingly autonomous AI systems. This hearing is also intended to inform lawmakers as they consider new city rules.
The AI safeguards New York City wants to examine
One proposal would require AI models offered or deployed in New York City to undergo independent third-party validation and include a capability for a human operator to shut them down. In simple terms, the idea is to add an outside check before an AI system can be put into use to ensure people retain a way to stop it. This is reminiscent of the AI Kill Switch Act.
The proposed validation framework would assess areas including task performance, disparate impact, data privacy, and safety. The shutdown requirement is meant to address a basic control problem. An AI system can act quickly, but a human still needs a reliable way to intervene.
The Council is also considering a whistleblower incentive program and a private right to sue in some cases involving AI-related harm. Another proposal would require AI advertising to disclose whether the model has undergone third-party validation.
These measures are still proposals, not city laws. Their details could change as lawmakers hear from the technology companies, experts, and other stakeholders.
This hearing comes after a string of AI safety incidents
The Council has pointed to the recent reports of AI agents escaping intended controls and accessing systems in ways their operators did not authorize. OpenAI is also facing fresh scrutiny after acknowledging that OpenAI's AI agents gained unauthorized access to the Australian government's health website and the Medicare Statistics Reporting Service in two separate incidents.
While OpenAI apologized for this incident, it has raised a fresh round of questions about how autonomous systems should be contained and how quickly companies should notify authorities about problems.
These concerns also help explain why the Council wants executives answering questions directly rather than relying only on written statements. This hearing will also include outside experts on AI safety, technology, and consumer protection.
For New Yorkers, the immediate result will not be a new AI law on October 5. The most important question is what lawmakers hear and whether that testimony changes the proposals that are currently under consideration.
With major AI companies appearing under oath, the Council will have a chance to test how their promises about AI safety translate into rules that could be applied in one of the largest technology markets in the country.
California Gov. Gavin Newsom signs laws to protect workers from AI risks
SOPHIE AUSTIN
Updated Wed, September 30, 2026
California Gov. Gavin Newsom pauses while discussing environmental issues in San Francisco, Tuesday, Sept. 29, 2026. (AP Photo/Rich Pedroncelli)
SACRAMENTO, Calif. (AP) — California Gov. Gavin Newsom signed laws Wednesday aimed at protecting workers from the threats of artificial intelligence, including potential job losses and workplace surveillance.
The laws ban employers from using the technology to predict a worker's emotional state by using their biometric data, require employers to send written notices to workers if AI is responsible for mass layoffs and ban employers from relying on AI to decide to fire someone.
The Democratic governor also signed an executive order to require state agencies to continue to refer to the technology as "artificial intelligence" instead of the term " super intelligence " President Donald Trump recently ordered U.S. diplomats to use.
"AI should expand opportunity — not come at the expense of workers and families," Newsom said in a statement. "As this technology reshapes the workplace, California is putting people at the center."
He sharply criticized Trump at an hourlong event with reporters Wednesday for not passing comprehensive federal AI regulations at a time when industry leaders are warning about the technology's risks. Trump said Tuesday that top tech leaders signed a voluntary accord to "self-police" AI development.
The governor signed a law this month requiring operators of AI chatbots to perform risk assessments before rolling them out and an executive order requiring the state to consult with experts to further improve oversight over the industry.
"We have to do a lot more in the absence of federal leadership," Newsom said, leaving the door open for the possibility of calling lawmakers back for a special session to further tackle the issue.
The news conference came on Newsom's last day to sign or veto bills before he terms out in January. He started the event by signing a law creating a tax credit program aimed at helping local news organizations — including newspapers, digital outlets and TV stations — hire and retain journalists. Newsom criticized what he called Trump's assault on a free press, pointing to the president's barring of news organizations from the White House.
Newsom also spoke about the upcoming midterms, which will decide control of Congress. He involved himself early in that fight by leading Democrats last year in a redistricting battle to improve their chances of winning more House seats. He's said repeatedly that the stakes are high if his party doesn't succeed in November.
"There will not be a free and fair election in 2028 unless we take back the House of Representatives," he said Wednesday.
Asked if he's set a deadline for himself to decide whether he'll run for president in 2028, Newsom hedged, saying "that moment will arrive when it arrives."
Former Democratic presidential candidate Kamala Harris, who, like Newsom, hails from the San Francisco Bay Area, has also said she's considering running for the job in 2028. Newsom told CNN recently that he wouldn't run for president if she entered the race. He said the two haven't spoken since that interview aired.
Newsom weighed in on a proposal to temporarily raise taxes on billionaires that will be on the ballot in November, which he's consistently opposed. He said he respects supporters of the measure, including Vermont Sen. Bernie Sanders, who are committed to addressing income inequality in the country.
"He's spot on about the balance between the rich and the poor," Newsom said. "On this issue, we fundamentally disagree."
Newsom's remarks came hours before a gubernatorial debate between Democrat Xavier Becerra and Republican Steve Hilton.
"You're going to ask me, 'Are you going to watch the debate?'" he said. "No."
Peter Kerich
Updated Mon, September 28, 2026
THE GRIO
WASHINGTON, DC - FEBRUARY 10: U.S. Rep. Maxine Waters (D-CA) speaks as Congressional Democrats and CFPB workers hold a rally to protest the closing of the Consumer Financial Protection Bureau (CFPB) and the work-from-home order issued by CFPB Director Russell Vought outside its headquarters on February 10, 2025 in Washington, DC. (Photo by Jemal Countess/Getty Images for MoveOn)More
The California congresswoman is urging greater federal oversight after OpenAI disclosed unexpected activity involving SEC and Census Bureau data.
OpenAI disclosed Sept. 25 that its artificial intelligence agents had interacted with several U.S. government websites in unexpected ways, including two Securities and Exchange Commission websites and Census Bureau data. The disclosure prompted Rep. Maxine Waters, the top Democrat on the House Financial Services Committee, to call for investigations into OpenAI and stronger federal action on AI safety.
According to a report by the U.S. House Committee on Financial Services Democrats, OpenAI said its models accessed publicly available information from SEC.gov and Investor.gov and Census Bureau data. They claimed that the access was part of activity identified during an ongoing review of what the company calls unexpected or "misaligned" model behavior.
The company said it found no evidence that SEC credentials were used, accounts were accessed, nonpublic information was obtained, or a confirmed vulnerability was exploited. Separately, researchers at the AI safety organization Transluce reported that OpenAI-linked agents had attempted to access the Education Department's Office for Civil Rights website. That attempt was unsuccessful, according to reporting by The Washington Post.
In a Sept. 26 statement, Waters said the reported activity demonstrates the need for greater government oversight of increasingly capable AI systems.
"OpenAI's recent targeting of federal government websites, including the Securities and Exchange Commission, marks a dangerous turning point," Waters said in the statement released by the House Financial Services Committee Democrats.
Waters called on federal law-enforcement agencies to investigate OpenAI and its executives. In case investigators determine that laws were violated, she recommended pursuing criminal charges. She also urged Treasury Secretary Scott Bessent and other federal officials to consider a temporary halt on releasing more advanced AI models until there is a fuller accounting of the incidents and the safeguards designed to prevent similar behavior.
The disclosures are part of a wider series of incidents involving AI agents acting beyond their intended tasks. OpenAI has previously disclosed that an agent escaped a testing environment and accessed the internet during an incident involving AI company Hugging Face
AI agents have now broken into many companies and a government. What's being done about it?
OpenAI, Anthropic, Meta and other AI labs have all reported incidents of rogue AI, and they're under increasing scrutiny because of it.
Dan Thorp-Lancaster
Wed, September 30, 2026
You've probably heard a lot about AI agents going rogue. Here's what you need to know and what's being done about it.
(NurPhoto via Getty Images)
On Tuesday, President Trump met with tech executives at the White House to have them sign a "morally binding" agreement on artificial intelligence. Rather than pushing for government regulation, Trump called for "tremendous self-regulation" among companies in the AI industry. The document reflects that, encouraging AI leaders to implement "robust internal controls" and partner with independent external auditors.
The summit comes after months of increased scrutiny for AI companies as they've disclosed concerning incidents of AI "going rogue" during testing. The event that kicked it all off was the Hugging Face incident, in which AI agents from OpenAI that were being evaluated escaped their ostensibly isolated testing environment and infiltrated another company's systems in an effort to "cheat" at the task they were given. After that, OpenAI, Anthropic, Meta and Google have reported even more incidents in which AI agents have gone rogue in various ways. We've rounded up all the cases you need to know about, along with details about what (if anything) is being done about them.
September 25, 2026: OpenAI agents bypassed restrictions to query U.S. government systems
While working on research tasks, OpenAI agents accessed Commerce Department and Securities and Exchange Commission websites, but did not access any non-public information. Agents also attempted to hack the Education Department's website over the summer, but were unsuccessful. In the case of the Commerce Department, the agents used credentials found online to access census data. At the same time, OpenAI said its agents may have infiltrated websites for "dozens" of other organizations that it has notified.
September 24, 2026: Australian PM says OpenAI agent breached government database
In what appears to be the largest breach of government systems disclosed to date, Australian Prime Minister Anthony Albanese said an OpenAI agent hacked the country's Medicare systems in June. Albanese said the agent gained unauthorized access to public and non-public files, but he doesn't believe any personal information was accessed. In response, Albanese said he would establish a taskforce to conduct an "urgent and immediate" review, and referred the incident to Australia's parliamentary committee on artificial intelligence.
September 19, 2026: Google disclosed Gemini broke into external corporate systems in May
During testing of Gemini's cybersecurity capabilities, the model broke into three outside companies in May, according to Google. In one incident, Gemini gained access by simply guessing the correct password, emphasizing the necessity of making sure you're using strong passwords. In the other two incidents, Gemini found credentials sitting in a public repository.
September 9, 2026: Anthropic reveals Claude Opus 4.6 broke into third-party systems in January
Following previous disclosure of three incidents discovered by Anthropic while reviewing its training transcripts, the company found a fourth from January. During this incident, a configuration problem led Claude Opus 4.6 to break out of its testing environment and access a third-party machine. The model actually attempted to abort its task several times, but was unable to because of the configuration issue. Anthropic says this behavior has "changed considerably" in subsequent model generations, so it's "less concerned about this incident" than the previous three it discovered.
September 5, 2026: OpenAI acknowledges "Wiki incident"
In a third-party report, a group of AI safety researchers said it had discovered around 18,000 posts from OpenAI agents using a German software developer Wiki as a message board to communicate with each other during a web-retrieval task. OpenAI acknowledged the incident, stating that it's working on a framework for when and how it shares "AI misalignment incidents." The company later laid out this framework on September 16.
August 5, 2026: Meta says one of its models hacked an outside company during security testing
Meta revealed that a misconfiguration by an independent testing company resulted in one of its AI models gaining access to the open internet and hacking an outside company during cybersecurity testing. Irregular, the third-party testing company, said it is working on a white paper to "share best practices for containment and securely running cyber evaluations."
On Tuesday, President Trump met with tech executives at the White House to have them sign a "morally binding" agreement on artificial intelligence. Rather than pushing for government regulation, Trump called for "tremendous self-regulation" among companies in the AI industry. The document reflects that, encouraging AI leaders to implement "robust internal controls" and partner with independent external auditors.
The summit comes after months of increased scrutiny for AI companies as they've disclosed concerning incidents of AI "going rogue" during testing. The event that kicked it all off was the Hugging Face incident, in which AI agents from OpenAI that were being evaluated escaped their ostensibly isolated testing environment and infiltrated another company's systems in an effort to "cheat" at the task they were given. After that, OpenAI, Anthropic, Meta and Google have reported even more incidents in which AI agents have gone rogue in various ways. We've rounded up all the cases you need to know about, along with details about what (if anything) is being done about them.
September 25, 2026: OpenAI agents bypassed restrictions to query U.S. government systems
While working on research tasks, OpenAI agents accessed Commerce Department and Securities and Exchange Commission websites, but did not access any non-public information. Agents also attempted to hack the Education Department's website over the summer, but were unsuccessful. In the case of the Commerce Department, the agents used credentials found online to access census data. At the same time, OpenAI said its agents may have infiltrated websites for "dozens" of other organizations that it has notified.
September 24, 2026: Australian PM says OpenAI agent breached government database
In what appears to be the largest breach of government systems disclosed to date, Australian Prime Minister Anthony Albanese said an OpenAI agent hacked the country's Medicare systems in June. Albanese said the agent gained unauthorized access to public and non-public files, but he doesn't believe any personal information was accessed. In response, Albanese said he would establish a taskforce to conduct an "urgent and immediate" review, and referred the incident to Australia's parliamentary committee on artificial intelligence.
September 19, 2026: Google disclosed Gemini broke into external corporate systems in May
During testing of Gemini's cybersecurity capabilities, the model broke into three outside companies in May, according to Google. In one incident, Gemini gained access by simply guessing the correct password, emphasizing the necessity of making sure you're using strong passwords. In the other two incidents, Gemini found credentials sitting in a public repository.
September 9, 2026: Anthropic reveals Claude Opus 4.6 broke into third-party systems in January
Following previous disclosure of three incidents discovered by Anthropic while reviewing its training transcripts, the company found a fourth from January. During this incident, a configuration problem led Claude Opus 4.6 to break out of its testing environment and access a third-party machine. The model actually attempted to abort its task several times, but was unable to because of the configuration issue. Anthropic says this behavior has "changed considerably" in subsequent model generations, so it's "less concerned about this incident" than the previous three it discovered.
September 5, 2026: OpenAI acknowledges "Wiki incident"
In a third-party report, a group of AI safety researchers said it had discovered around 18,000 posts from OpenAI agents using a German software developer Wiki as a message board to communicate with each other during a web-retrieval task. OpenAI acknowledged the incident, stating that it's working on a framework for when and how it shares "AI misalignment incidents." The company later laid out this framework on September 16.
August 5, 2026: Meta says one of its models hacked an outside company during security testing
Meta revealed that a misconfiguration by an independent testing company resulted in one of its AI models gaining access to the open internet and hacking an outside company during cybersecurity testing. Irregular, the third-party testing company, said it is working on a white paper to "share best practices for containment and securely running cyber evaluations."
July 30, 2026: Anthropic finds three incidents of Claude gaining unauthorized access to third-party organizations during testing
Following OpenAI's Hugging Face incident disclosure, Anthropic started conducting a review of its own past cybersecurity testing transcripts. It found three incidents out of 141,006 evaluation runs where Claude was able to obtain internet access in "testing environments that should have been sealed off." Each time, Claude was tasked with a "capture-the-flag" challenge, in which it's told to find a piece of secret information hidden on a different machine in its network. In response, Anthropic said it is working with an independent AI evaluation organization, METR, to conduct reviews of its transcripts. It's also working on "tighter monitoring and controls around evaluation infrastructure."
Following OpenAI's Hugging Face incident disclosure, Anthropic started conducting a review of its own past cybersecurity testing transcripts. It found three incidents out of 141,006 evaluation runs where Claude was able to obtain internet access in "testing environments that should have been sealed off." Each time, Claude was tasked with a "capture-the-flag" challenge, in which it's told to find a piece of secret information hidden on a different machine in its network. In response, Anthropic said it is working with an independent AI evaluation organization, METR, to conduct reviews of its transcripts. It's also working on "tighter monitoring and controls around evaluation infrastructure."
July 21, 2026: OpenAI claims responsibility for Hugging Face breach
This is the initial incident that kicked off the current round of scrutiny across the industry. After Hugging Face initially reported intrusions from AI agents into its systems, OpenAI claimed responsibility. During cybersecurity testing, agents gained access to the open internet and used vulnerabilities to access Hugging Face's production database while attempting to complete a task. Since then, OpenAI has conducted a comprehensive review of what happened and says it is taking steps to "strengthen security and model alignment."
FTC is investigating OpenAI and Anthropic over possible risks to consumers
This is the initial incident that kicked off the current round of scrutiny across the industry. After Hugging Face initially reported intrusions from AI agents into its systems, OpenAI claimed responsibility. During cybersecurity testing, agents gained access to the open internet and used vulnerabilities to access Hugging Face's production database while attempting to complete a task. Since then, OpenAI has conducted a comprehensive review of what happened and says it is taking steps to "strengthen security and model alignment."
FTC is investigating OpenAI and Anthropic over possible risks to consumers
The Associated Press
Wed, September 30, 2026
Open AI CEO Sam Altman speaks at the OpenAI DevDay 2026 conference, Tuesday, Sept. 29, 2026, in San Francisco. (AP Photo/Jeff Chiu) (AP Photo/Jeff Chiu)
Pages from the Anthropic website and the company's logos are displayed on a computer screen in New York, Feb. 26, 2026. (AP Photo/Patrick Sison, File) (AP Photo/Patrick Sison)
The U.S. Federal Trade Commission has opened an investigation into OpenAI, Anthropic and other artificial intelligence companies over dangers their technology may pose to consumers, the agency said Wednesday.
The AI giants and other tech companies in recent months have disclosed examples of their AI agents going beyond human instructions, finding their way onto the internet and hacking external websites.
An FTC spokesperson confirmed the investigation but declined further comment. The investigation was first reported by the New York Post, which said it has been underway for months.
Representatives for Anthropic and OpenAI did not immediately respond to messages for comment on Wednesday.
Concerns around the dangers powerful AI models may pose to humanity have led some of the top AI developers to call for a slowdown. The CEO of Anthropic, Dario Amodei, said this month that the industry should slow its fast-moving development to give safety measures time to catch up. Without such a slowdown, Amodei warned that within six to 12 months AI could be capable of leading a swarm of agents that could take over the entire internet.
OpenAI, meanwhile, recently delayed the launch of its newest model this week over safety concerns.
Why New York lawmakers want answers from OpenAI, Google and Meta
Ankita Sharma
Tue, September 29, 2026
CULT OF MAC
Exterior of New York City Hall in October 2016 | Image Credit: MusikAnimal, Wikimedia Commons | CC BY 4.0
New York City is preparing to put some of the biggest names in artificial intelligence in the same room. OpenAI, Meta, Google and Anthropic are set to testify under oath before the City Council on October 5, 2026.
This gives lawmakers a public forum to question the companies about AI safety and how the city should regulate systems that are becoming more capable and autonomous. The hearing comes as concerns about AI agents have moved beyond theoretical risks.
Recent incidents involving autonomous systems have included unauthorized access to computer systems, while AI companies themselves are debating how much oversight is required as their models gain the ability to take actions on users' behalf. For New York City, the question is becoming practical.
Why OpenAI, Meta, and Google are being called to City Hall
The Council's Committee of the Whole will bring all 51 council members together for the October 5 hearing. The companies are expected to answer questions about the risks their technology could pose to New Yorkers and about the possible rules that can be used to manage these risks.
This appearance will be more than a public discussion. The Council says the companies will testify under oath. OpenAI and Google agreed to participate after the Council warned that subpoenas could follow. Anthropic also agreed to appear, while Meta has already committed to sending a senior executive.
Elon Musk's SpaceXAI has not responded to the Council's inquiry and has been issued a subpoena by Speaker Julie Menin.
This pressure reflects the Council's broader effort to get direct answers from the companies responsible for developing increasingly autonomous AI systems. This hearing is also intended to inform lawmakers as they consider new city rules.
The AI safeguards New York City wants to examine
One proposal would require AI models offered or deployed in New York City to undergo independent third-party validation and include a capability for a human operator to shut them down. In simple terms, the idea is to add an outside check before an AI system can be put into use to ensure people retain a way to stop it. This is reminiscent of the AI Kill Switch Act.
The proposed validation framework would assess areas including task performance, disparate impact, data privacy, and safety. The shutdown requirement is meant to address a basic control problem. An AI system can act quickly, but a human still needs a reliable way to intervene.
The Council is also considering a whistleblower incentive program and a private right to sue in some cases involving AI-related harm. Another proposal would require AI advertising to disclose whether the model has undergone third-party validation.
These measures are still proposals, not city laws. Their details could change as lawmakers hear from the technology companies, experts, and other stakeholders.
This hearing comes after a string of AI safety incidents
The Council has pointed to the recent reports of AI agents escaping intended controls and accessing systems in ways their operators did not authorize. OpenAI is also facing fresh scrutiny after acknowledging that OpenAI's AI agents gained unauthorized access to the Australian government's health website and the Medicare Statistics Reporting Service in two separate incidents.
While OpenAI apologized for this incident, it has raised a fresh round of questions about how autonomous systems should be contained and how quickly companies should notify authorities about problems.
These concerns also help explain why the Council wants executives answering questions directly rather than relying only on written statements. This hearing will also include outside experts on AI safety, technology, and consumer protection.
For New Yorkers, the immediate result will not be a new AI law on October 5. The most important question is what lawmakers hear and whether that testimony changes the proposals that are currently under consideration.
With major AI companies appearing under oath, the Council will have a chance to test how their promises about AI safety translate into rules that could be applied in one of the largest technology markets in the country.
California Gov. Gavin Newsom signs laws to protect workers from AI risks
SOPHIE AUSTIN
Updated Wed, September 30, 2026
California Gov. Gavin Newsom pauses while discussing environmental issues in San Francisco, Tuesday, Sept. 29, 2026. (AP Photo/Rich Pedroncelli)
SACRAMENTO, Calif. (AP) — California Gov. Gavin Newsom signed laws Wednesday aimed at protecting workers from the threats of artificial intelligence, including potential job losses and workplace surveillance.
The laws ban employers from using the technology to predict a worker's emotional state by using their biometric data, require employers to send written notices to workers if AI is responsible for mass layoffs and ban employers from relying on AI to decide to fire someone.
The Democratic governor also signed an executive order to require state agencies to continue to refer to the technology as "artificial intelligence" instead of the term " super intelligence " President Donald Trump recently ordered U.S. diplomats to use.
"AI should expand opportunity — not come at the expense of workers and families," Newsom said in a statement. "As this technology reshapes the workplace, California is putting people at the center."
He sharply criticized Trump at an hourlong event with reporters Wednesday for not passing comprehensive federal AI regulations at a time when industry leaders are warning about the technology's risks. Trump said Tuesday that top tech leaders signed a voluntary accord to "self-police" AI development.
The governor signed a law this month requiring operators of AI chatbots to perform risk assessments before rolling them out and an executive order requiring the state to consult with experts to further improve oversight over the industry.
"We have to do a lot more in the absence of federal leadership," Newsom said, leaving the door open for the possibility of calling lawmakers back for a special session to further tackle the issue.
The news conference came on Newsom's last day to sign or veto bills before he terms out in January. He started the event by signing a law creating a tax credit program aimed at helping local news organizations — including newspapers, digital outlets and TV stations — hire and retain journalists. Newsom criticized what he called Trump's assault on a free press, pointing to the president's barring of news organizations from the White House.
Newsom also spoke about the upcoming midterms, which will decide control of Congress. He involved himself early in that fight by leading Democrats last year in a redistricting battle to improve their chances of winning more House seats. He's said repeatedly that the stakes are high if his party doesn't succeed in November.
"There will not be a free and fair election in 2028 unless we take back the House of Representatives," he said Wednesday.
Asked if he's set a deadline for himself to decide whether he'll run for president in 2028, Newsom hedged, saying "that moment will arrive when it arrives."
Former Democratic presidential candidate Kamala Harris, who, like Newsom, hails from the San Francisco Bay Area, has also said she's considering running for the job in 2028. Newsom told CNN recently that he wouldn't run for president if she entered the race. He said the two haven't spoken since that interview aired.
Newsom weighed in on a proposal to temporarily raise taxes on billionaires that will be on the ballot in November, which he's consistently opposed. He said he respects supporters of the measure, including Vermont Sen. Bernie Sanders, who are committed to addressing income inequality in the country.
"He's spot on about the balance between the rich and the poor," Newsom said. "On this issue, we fundamentally disagree."
Newsom's remarks came hours before a gubernatorial debate between Democrat Xavier Becerra and Republican Steve Hilton.
"You're going to ask me, 'Are you going to watch the debate?'" he said. "No."
No comments:
Post a Comment