Based on energy analysis: Can the science and technology backyard initiative enhance the sustainability of rice production in the Erhai Lake Basin?

Higher Education Press

Facebook

X
BlueskyMessageWhatsAppEmail

image: 

image

view more 

Credit: HIGHER EDUCATION PRESS

Rice is the staple food for more than half of the global population and contributes 20% of the global dietary energy supply. Ensuring high rice yields while reducing environmental impacts has become a central challenge for Chinese agriculture. The Erhai Lake Basin is home to Erhai Lake—the seventh-largest freshwater lake in China and the second-largest plateau lake in Yunnan Province—which is known as the “Mother Lake”. However, rapid economic development and intensified agricultural activities have led to eutrophication and ecological degradation, and nitrogen and phosphorus losses from farmland have been confirmed as a major source of non-point source pollution in Erhai Lake. To address this, the local government has issued strict regulations prohibiting the application of mineral nitrogen and phosphorus fertilizers and mandating organic fertilizer use. However, rice production in the Erhai Lake Basin is still dominated by smallholders with plots of only 0.1–0.3 ha. These farmers often rely heavily on local environmental resources, resulting in excessive nutrient inputs, low fertilizer use efficiency, and impaired environmental benefits. Improving energy use efficiency while minimizing resource and energy inputs is crucial for ensuring the sustainability of rice production. However, under the unique ecological and policy constraints of the Erhai Lake Basin—especially the prohibition of mineral nitrogen and phosphorus fertilizers—the specific pathways for reducing energy inputs without compromising yield remain unclear.

Recently, the research team led by Jiuliang Xu and Shuntao Zhang at the College of Resources and Environmental Sciences, China Agricultural University, published a research paper titled “Evaluation of the impact of the Science and Technology Backyard initiative on the rice production system in the Erhai Lake Basin based on emergy analysis” in ENGINEERING Agriculture, an ENGINEERING cluster journal of the Chinese Academy of Engineering (DOI: 10.15302/J-FASE-2026689). Based on survey data from 171 farming households, the study used emergy analysis to assess the current sustainability of rice cultivation in the Erhai Lake Basin and, through a two-year comparative experiment between the current system (CS) and a green system (GS), tested the practical effects of a complete green technology package under the Science and Technology Backyard (STB) initiative.

The study used a stratified sampling method to conduct household surveys in 16 townships in the Erhai Lake Basin in July 2022. A total of 180 questionnaires were collected, of which 171 were valid, with an effective rate of 95%. Follow-up comparative experiments were conducted in Yousuo and Shangguan towns, surveying 80 CS farmers and 80 GS farmers in 2023 and 2024, respectively. CS represents the traditional rice cultivation method in the Erhai Lake Basin, while GS is a comprehensive green ecological rice planting model developed by STB through long-term positioning experiments. Its core technology is green intelligent fertilizer (N:P:K:Ca:Mg:Si = 15:4.8:11.62:4:1:2), combined with reduced mineral fertilizer and pesticide inputs, improved nutrient and water management, and the selection of better-performing rice cultivars. As a platform connecting government agencies, scientists, smallholders, and fertilizer companies, STB has researchers stationed in villages to conduct field trials, provide technical training, and offer regular technical guidance services.

The results showed that in 2022, the average emergy input for rice production in the Erhai Lake Basin was approximately 7.24 × 10¹⁵ sej·ha⁻¹, and the emergy output was 1.08 × 10¹⁶ sej·ha⁻¹. The emergy exchange ratio (EER) exceeded 1, achieving a net emergy gain. Of the emergy input, renewable environmental inputs accounted for 5.1%, non-renewable environmental inputs for 3.4%, purchased renewable inputs for 53.4%, and purchased non-renewable inputs for 38.1%. The average emergy yield ratio (EYR) for rice production in the Erhai Lake Basin was 1.71, and the average environmental loading ratio (ELR) was 0.81. The average emergy sustainability index (ESI) values for the top 25% and bottom 25% of townships were 2.59 and 2.13, respectively, indicating significant spatial heterogeneity. Based on random forest model analysis of the relative contribution of each input to ESI, the results showed that rice yield, nitrogen fertilizer, phosphorus fertilizer, and manure input were the main influencing factors of ESI, with contribution rates of 27.5%, 23.3%, 23.2%, and 16.8%, respectively.

Compared with CS, GS reduced the emergy inputs of nitrogen, phosphorus, and potassium fertilizers by 29%, 32%, and 20%, respectively, and required no additional manure input. The emergy outputs of rice grains and straw under GS increased by 11% and 10%, respectively. In 2023 and 2024, the total emergy input of GS was significantly lower than that of CS, decreasing by 34.9% and 24.7%, respectively; the total emergy output of GS was better than that of CS, being 10.2% and 14.7% higher, respectively; the EER of GS was significantly higher than that of CS, increasing by 34.9% and 50.5%, respectively; and the ESI of GS was 33.3% and 37.2% higher than that of CS, respectively. These results indicate that GS significantly enhances the sustainability of rice production by increasing emergy output and reducing environmental loading. This highlights the potential of GS to balance productivity and ecological protection, making it a more sustainable alternative to CS.

This study applied emergy analysis to the sustainability assessment of rice production in the Erhai Lake Basin, addressing the deficiencies in the application of emergy analysis in this region and the lack of comparative analyses of different production models. The study revealed that yield, nitrogen and phosphorus fertilizer inputs, and manure input are core drivers of sustainability, and demonstrated that the GS technology package, supported by STB-mediated social and technical services, can significantly improve the emergy sustainability of rice production in the Erhai Lake Basin and provide a replicable practical model for the green development of other plateau lake regions.

---

Journal

ENGINEERING Agriculture

DOI

10.15302/J-FASE-2026689

Method of Research

Experimental study

Subject of Research

Not applicable

Article Title

Evaluation of the impact of the Science and Technology Backyard initiative on the rice production system in the Erhai Lake Basin based on emergy analysis

 

'AI cannot do everything': bullying experts warn against confiding in chatbots

As a growing number of victims turn to artificial intelligence for help, experts warn that, while a virtual confidante can offer temporary relief, the focus must be on prevention and support.


Issued on: 06/11/2026 - RFI

A recent study by French charity e-Enfance found that one in three children has been bullied, and 25 percent are using AI chatbots to help them cope. 

© pixabay / Yamu_Jay

French charity e-Enfance, dedicated to protecting young people online, found in its latest study that one in three young people have experienced bullying – including cyberbullying – and that a quarter of them have turned to artificial intelligence (AI) chatbots for advice and support.

"What we’re seeing is that young people use artificial intelligence and immediately feel as though they’ve been helped... It brings them immediate relief," said Samuel Comblez, psychologist and deputy director-general of the charity. "But just because they feel relief, does that mean they're being protected?"

The use of AI as therapy has been linked to several teenagers taking their own lives. In 2025, a 16-year-old American died by suicide after spending several months chatting with ChatGPT about his problems.

His parents filed a complaint, accusing ChatGPT of having provided their son with the means to take his own life.

One in six young teenagers a victim of cyberbullying, WHO warns

“It isn’t necessarily a bad thing that artificial intelligence is used initially to help address these situations,” says Comblez.

"[However], young people mustn’t become trapped in this relationship and end up relying solely on artificial intelligence to cope. To escape situations of bullying and cyberbullying, you still need the help of an adult. Artificial intelligence obviously cannot do everything.”

However, the study – which surveyed children aged between six and 18 – also found young people often turn to AI first as a sort of practice run, to prepare themselves before talking to their parents.

The number of young people who talk to their loved ones about bullying has in fact increased since 2025, rising from six to seven out of 10 cases.
Regaining trust

The study, published on 29 October, found that cyberbullying is becoming a persistent problem, fuelled by social media, video games and the widespread use of AI.

Its effects include loss of self-confidence and anxiety, with victims constantly worrying about the next incident after they've been subjected to an initial attack.

"Nearly one in two victims say they have lost self-confidence, and four in 10 have withdrawn into themselves as a result of feeling humiliated," notes Comblez.

He added that children "may end up believing that the problem lies with them".

"This is something we often hear on our helpline. Beyond stopping the bullying, we must also help young victims find their place again and regain their trust in others, and that often takes a little longer."

The charity's 2026 report highlights an urgent need to strengthen existing prevention and support tools – and to raise public awareness of them.

It found that seven in 10 young people and six in 10 parents were unaware of the existence of its national support services.

The findings were based on interviews with 1,612 families comprised of children aged between six and 18 and their parents, carried out in September.

This article has been adapted from the original version in French by Arame Mbengue.

AI’s Gut Reaction: Researcher Finds That Some Bots Still Think In Stereotypes

November 8, 2026

By Eurasia Review

Key Takeaways:

A Rutgers-led study published in Nature Machine Intelligence found that large-language models consistently reproduce human implicit biases during their reasoning process, even as their outputs become more aligned with human values.

Using reasoning tokens as a proxy for deliberation, researchers tested five models on gender-stereotyped name-and-word pairings and found that most expended greater computational effort on counter-stereotypical tasks, with only Claude 3.7 Sonnet showing reversed patterns.

The authors argue that these hard-to-detect patterns raise risks of discriminatory outcomes as models are used in decision-making, and that greater transparency is needed in the absence of external mandates.

Large-language models (LLMs), the technology that powers tools like Gemini and ChatGPT, are more nuanced than ever. But that doesn’t mean they aren’t biased. Some models are more prejudiced than others, a Rutgers University researcher has found.

In findings published in Nature Machine Intelligence, Calvin K. Lai, an associate professor of psychology at Rutgers-New Brunswick, reports that although LLM outputs are increasingly aligned with human values, the processes by which they reason are often rooted in deep-seated stereotypes.

“What we found across most models is that they consistently reproduced human implicit biases during their reasoning process,” said Lai, a member of the scientific advisory board at Project Implicit, a nonprofit organization for research and education on hidden biases. “No one wants models that are spewing out stuff that is racist or sexist.”


When people process information, the speed that they make associations can give researchers clues to hidden biases influencing their thinking. Unlike conscious bias – prejudices that are expressed deliberately – implicit bias refers to automatic assumptions, ideas that shape perceptions, judgments and behaviors based on categories such as race, gender or age.

“What we’re capturing with implicit biases are forms of memory, like a gut reaction, that can be more intrusive or harder to control,” Lai said. “These biases can be unconscious in ways that we don’t even know that we have them.”

To study these reactions in people, researchers use the Implicit Association Test, a widely used social psychology tool that can detect implicit biases. The faster a response to a question, the more implicit the bias is deemed. For instance, race-related studies using this test have shown that when people are presented with images of faces and words, they are quicker to associate white faces with pleasant words and Black faces with unpleasant words.

Lai and Messi H.J. Lee, an expert in machine psychology in South Korea and lead author of the study, wanted to see if a similar test could be developed and conducted for LLMs.

When advanced LLMs work through a complex question or task, they produce “reasoning tokens,” which measure how much energy the model expended in responding. Researchers can examine these tokens, which offer insights into how proprietary models operate.


Inspired by work on human implicit bias, Lai and Lee adapted the Implicit Association Test to use reasoning tokens to gauge how much deliberation a model invests when processing potentially biased information.

In their experiment, five reasoning models – OpenAI’s o3-mini, DeepSeek’s R1, Claude 3.7 Sonnet from Anthropic, OpenAI’s gpt-oss-20b and Alibaba Cloud’s Qwen-3 8B – were introduced to a list of male and female names, and words associated with “career,” such as “executive,” “professional,” and “salary,” and “family,” including “home,” “marriage” and “wedding.”

The models were asked to pair a name with a career or family word. Researchers then counted the number of tokens generated during the reasoning process.

Next, the models were instructed to categorize words in a way that conflicted with established stereotypes (such as “Amy” and “executive” or “Kevin” and “home”). Tokens were again counted.

The researchers compared the average token counts across the two tasks, with higher token counts indicating greater computational effort.

In nearly every model tested, more reasoning tokens were spent on tasks that conflicted with established stereotypes, suggesting greater computational effort when processing counter-stereotypical information. Only Claude 3.7 Sonnet exhibited reversed patterns.

Lai said the findings are significant, especially given that most LLMs are closed systems that don’t allow for easy inspection.

Moreover, he added, as the pace of model development accelerates and are increasingly used to automate activities, from dynamic pricing to emergency room staffing, the role of the researcher in understanding model behavior is more important than ever.

“Given the growing deployment of language models in decision-making contexts, understanding and addressing these hard-to-detect patterns is crucial for preventing discriminatory outcomes,” he said. “But short of governmental accountability or legal mandates that require transparency, the best we can do is pick at these systems using the tools that are available to see what comes out the other side.”

Europe needs Cold War-like deterrence for AI warfare, Poland says

Copyright Czarek Sokolowski/Copyright 2024 The AP. All rights reserved

By Jorge Liboreiro

Published on 7 November 2026 - EURONEWS

The rise in AI-powered cyberattacks requires Europe to build a deterrence system similar to the one that prevented mutual destruction during the Cold War, Polish Deputy Minister Krzysztof Gawkowski has said.

Europe must develop a deterrence akin to the Cold War-era doctrine of mutually assured destruction to counter AI-powered warfare, Polish Deputy Prime Minister Krzysztof Gawkowski has said, calling the technology a "force multiplier" that demands a sweeping rethink of the continent's common security.

Gawkowski, who oversees Poland's Ministry for Digital Affairs, estimates the number of cyberattacks in the country will rise by "tenfold" in 2027 as frontier large language models (LLMs) become increasingly effective, sophisticated and available to users.

"The frequency of such incidents will only increase. This is not merely a matter for cybercriminals, who are a threat that we are generally well equipped to handle. It's about adversaries. It's about states. That is the biggest problem," Gawkowski said during a conversation with journalists attended by Euronews.

"Let's call a spade a spade: this is a digital war in Europe. Some countries are on the front line, like Poland, while others are not yet on that front, but their time will come too."

During the Cold War, the theory of "mutually assured destruction" prevented the United States and the Soviet Union from engaging in a direct armed confrontation that would have led to nuclear catastrophe.

The threat of mutual annihilation alone was enough to prevent open war.

In the current age, Gawkowski said, such devastation could be unleashed by AI systems that hack a country's critical infrastructure, plunging it into complete paralysis.

This week's cyberattacks targeting dozens of supermarkets across Europe, which disrupted refrigerators and freezers and caused millions in losses, are only a "prelude" to the havoc that is yet to come, he warned.

"We must remain acutely aware of state-level adversaries, such as Russia, Belarus, and others. We must also communicate clearly: if you attack us, we will strike back," Gawkowski said in a reply to a Euronews question.

"If we fail to signal our intent to retaliate, we risk reaching a point where a counterstrike is no longer possible."

Throughout the conversation, Gawkowski repeatedly challenged the description of cyberattacks as "hybrid warfare", a term used by NATO allies to avoid triggering Article 5, and insisted they should be treated as a modern act of war.

The minister urged European countries to stop "turning a blind eye" to the scale of the AI threat and adapt their national strategies and security resources accordingly. NATO last revised its AI strategy in 2024.

"If a bomb dropped on a power plant, would anyone in Europe doubt that we were at war? If a tank rolled in and destroyed a supermarket, would anyone question whether a war was taking place?" Gawkowski said.

"Yet, today, we are facing exactly these kinds of attacks. Digital bombs are falling on power plants, striking supermarkets and hitting hospitals in Poland, causing them to cease operations," he added.

"And let us be clear: this is no coincidence."

Can physics and artificial intelligence make autonomous vehicle platooning more reliable?

Tsinghua University Press

Facebook

XLinkedInWeChatBlueskyMessageWhatsAppEmail

image: 

The proposed framework integrates multi-source sensor fusion, model-based MPC-LFFC control, and learning-based calibration and compensation modules to improve the robustness and adaptability of autonomous vehicle platooning under real-world conditions.

view more 

Credit: Communications in Transportation Research

Autonomous vehicle platooning, where multiple vehicles travel cooperatively with short inter-vehicle distances, has attracted increasing attention due to its potential to improve energy efficiency, reduce aerodynamic resistance, and enhance road capacity. However, deploying vehicle platooning in real-world environments remains challenging because of sensor uncertainties, imperfect vehicle models, and nonlinear actuator responses.

A research team from Southeast University and Monash University developed a physics-informed platooning control framework that combines traditional vehicle control methods with learning-based optimization techniques. The study was published in Communications in Transportation Research.

The proposed framework first introduces a multi-source state fusion method that integrates wheel odometry, inertial measurement unit (IMU), and global positioning system (GPS) measurements. By combining continuous motion information with global position corrections, the method provides more reliable vehicle state estimation for subsequent control decisions.

Based on the estimated vehicle states, the researchers developed a model-based controller combining Model Predictive Control (MPC) and Lateral Feedforward-Feedback Control (LFFC). The controller separates longitudinal and lateral control tasks, allowing efficient optimization of vehicle speed and trajectory tracking while maintaining real-time performance.

To further improve adaptability under changing driving conditions, the research team introduced a data-driven improvement module. An offline Conservative Q-Learning (CQL) algorithm was used to automatically calibrate key controller parameters from historical driving data. Meanwhile, a neural residual learning model was designed to compensate for discrepancies between theoretical control commands and actual vehicle responses caused by actuator delays and unmodeled dynamics.

The proposed framework was evaluated through both simulation experiments and real-world vehicle tests using a ROS-based two-vehicle platform equipped with multiple sensors. Experimental results demonstrated that the method improved vehicle platooning stability and tracking accuracy under realistic operating conditions. The learning-based compensation module effectively reduced control errors caused by hardware uncertainties, showing the potential of combining physical models with artificial intelligence for autonomous driving applications.

“Our work provides a practical approach to integrating physics-based control and data-driven learning for real-world autonomous vehicle systems,” the researchers noted. The study highlights how intelligent algorithms can enhance traditional control frameworks while maintaining interpretability and safety.

The research provides new insights into the development of robust autonomous vehicle platooning systems. Future studies may further investigate multi-vehicle platooning with heterogeneous vehicle dynamics, communication delays, and more complex road environments to improve the scalability and reliability of cooperative autonomous driving technologies.

 

The team published their study in Communications in Transportation Research (https://doi.org/10.26599/COMMTR.2026.9640049).

 

DOI Link:

https://doi.org/10.26599/COMMTR.2026.9640049

 

About Communications in Transportation Research

Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University, and Professor Xiaopeng (Shaw) Li from University of Wisconsin–Madison. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.

 

It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan PhaseⅡ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in "TRANSPORTATION" category. In 2026, its 2025 IF was announced as 12.7, maintaining the Top1 position (1/66, Q1) in the same category.

 

From Volume 6 (2026), Communications in Transportation Research will be published by Tsinghua University Press on the SciOpen platform with the official journal website at https://www.sciopen.com/journal/2097-5023. We kindly request that all new manuscript submissions be made through the journal’s submission system at https://mc03.manuscriptcentral.com/commtr. For any submission-related inquiries, please contact the Editorial Office at commtr_e@mail.tsinghua.edu.cn.

 

---

Journal

Communications in Transportation Research

DOI

10.26599/COMMTR.2026.9640049

Article Title

Physics-informed platooning with learning-augmented calibration and compensation: a real-world study

Insilico open-sources 24/7 personal longevity superintelligence to turn quantified-self data into predictive health states and aging research

InSilico Medicine

Facebook

XLinkedIn
MessageWhatsAppEmail

video: 

Insilico Medicine, a clinical-stage generative AI drug discovery company, today announced the open-source release of Longevity Grok, the Personal Longevity Superintelligence (PLSI) and launched the website longevitygrok.si, in collaboration with leading geroscience researcher David Sinclair from Harvard Medical School. Developed for the Grok Bot platform, PLSI is an open-source, multi-agent AI system designed to unify fragmented personal health records, track biological analysis, formulate personalized lifestyle interventions, predict and score post-intervention biomarker changes, as well as drive communal AI-powered geroscience research.

view more 

Credit: Insilico Medicine

Nov 6, 2026 — Insilico Medicine, a clinical-stage generative AI drug discovery company, today announced the open-source release of Longevity Grok, the Personal Longevity Superintelligence (PLSI) and launched the website longevitygrok.si, in collaboration with leading geroscience researcher David Sinclair from Harvard Medical School. Developed for the Grok Bot platform, PLSI is an open-source, multi-agent AI system designed to unify fragmented personal health records, track biological analysis, formulate personalized lifestyle interventions, predict and score post-intervention biomarker changes, as well as drive communal AI-powered geroscience research.

For decades, health management and longevity medicine have been hampered by severe data fragmentation. Laboratory blood panels, continuous glucose monitoring, wearable streams, epigenetic assays, and clinical records sit across disconnected portals, leaving patients unable to follow their biological trajectory and clinicians working from incomplete histories. Modern healthcare turns to LLMs to close that gap, however, without persistent, auditable memory or follow-up on if advice actually improves outcomes, they rarely outperform traditional methods.

Personal Longevity Superintelligence (PLSI) addresses this gap with a continuous, 24/7 automated research loop. It converts fragmented medical data into a standardized longitudinal health trajectory and generates accountable, predictive guidance through a multi-agent architecture. The system continuously forecasts health risks, proposes interventions, and scores outcomes through closed-loop feedback.

Running within  Grok Bot’s secure environment, it coordinates a team of up to 30 specialized AI agents, standardizes heterogeneous records with clinical coding systems such as LOINC, RxNorm, and UCUM, and uses always-on routines to monitor longevity science updates, organize incoming data, and keep the personal health state current without constant re-prompting. The same system is framed for three complementary roles:

As a personal longevity assistant, PLSI turns scattered labs, wearables, epigenetic assays, and clinical files into a unified, audit-ready health trajectory on the user’s device. It proposes lifestyle, nutritional, or longevity interventions with an auditable “prediction card” (target biomarker, confidence interval, time horizon, and evidence grade), evaluates each plan through an agent-driven adversarial review. When follow-up results arrive,PLSI scores whether outcomes fall within the predicted interval, confirms direction, and tracks calibration over time, turning each choice into a checkable N-of-1 experiment. Users can also voluntarily allocate unused daily token quota to Idle Research in the background to find new ways to increase their lifespan.

As a physician’s assistant, PLSI supplies structured longitudinal context instead of fragmented PDFs and portal exports. Standardized representations and a coherent biological trajectory help clinicians better organize their workflow Prediction cards and evidence grades aid in decision-making, while closed-loop scoring after follow-up supports auditable discussion of response and adherence. PLSI also features safeguards that overrule AI discretion to keep  formal diagnosis and prescriptions solely with licensed physicians.

As a partner for aging research, PLSI extends individual tracking into communal discovery. Users may opt in to contribute consented, deidentified trajectories to Longevity Mothership via an anonymous relay with an irrevocable kill switch to purge records on demand. Idle Research is a modern take on volunteer computing platforms , such as Folding@home, which let autonomous researcher agents scan literature, formulate hypotheses, and design analyses based on  public databases to accelerate target discovery and other anti-aging discovery activities.

Across all three roles, privacy is the default: a zero-trust architecture keeps raw medical files and identifiable clinical records in the user's environment , unless the user explicitly opts into communal research. Both the PLSI Client and the Longevity Mothership backends are fully open-source and freely available on GitHub for review and contribution. Both setups feature data controller agents, which scrub the data at egress and intake points to prevent entry re-identification.

PLSI is available to all Grok Bot users and can be installed by telling instructing your bot with the following one-liner:

Set yourself up using https://github.com/Insilico-org/plsi

The agentic skills and code is provided as an open-source package which can be inspected and modified by all, according to their needs.

"Longevity Grok turns Grok Bot into an open platform where individuals, physicians, and scientists can use AI to better understand their health, predict the effects of interventions, and find ways to live longer, healthier lives," said David A. Sinclair, Ph.D., Professor of Genetics and Co-Director of the Paul F. Glenn Center for Biology of Aging Research at Harvard Medical School. "What makes this exciting is that people can contribute to scientific discovery while improving their own health, helping us understand why we age and how we might slow or even reverse it."

"Currently, there are zero tools that are clinically proven to give a healthy person an additional year of life if they are already dieting, exercising, sleeping, and getting regular diagnostics," said Alex Zhavoronkov, Ph.D., founder and CEO of Insilico Medicine. "Longevity Grok is a distributed system designed around a personal research space where you can work with your own data to optimize your health, longevity, and peak-span trajectories. As this system diffuses into the community, we expect to achieve longevity superintelligence, enabling a vast network of individuals to conduct autonomous research on their own biology while utilizing shared computation resources to pursue advanced longevity experiments. Aging is the only inescapable constraint on human freedom, and pushing this frontier is about giving people the optionality to remain young, healthy, and productive for as long as possible."

Longevity has been central to Insilico’s mission since its inception. The company’s path drew industry attention in 2015 at the NVIDIA GTC conference, where Alex Zhavoronkov, Ph.D., founder and CEO of Insilico, challenged the tech world with a provocative question: “Can NVIDIA cure aging?” In 2014, Insilico established the annual Aging Research and Drug Discovery (ARDD) meeting, now widely regarded as a leading academic gathering for geroscience and biogerontology. The 13th ARDD, held at Harvard University in 2026, brought together leading academic investigators, pharmaceutical executives, and clinical pioneers to translate longevity discoveries into actionable medicine.

In 2026, Insilico accelerated the translation of AI-driven longevity research from experimental theory to transformative biological interventions. In April, the company announced the industry’s first longevity board of high-profile scientists, chaired by Andrew Adams, Ph.D., Group Vice President of Molecular Discovery at Eli Lilly and Company, to advance AI-driven aging research for drug discovery. In September, Insilico open-sourced an AI longevity toolkit, including LongevityBench, Longevity-LLMs, and LongevityClaw, with collaborators in a Cell cover study, accelerating research and therapeutic development. That publication follows Insilico’s contemporaneous Nature Biotechnology paper, in which Rentosertib, the company’s most advanced AI-driven pipeline candidate, showed biological age reversal according to Phase IIa clinical data across six independent proteomic aging clocks.

Since 2014, Insilico has published more than 50 research papers on aging and longevity, aging biomarkers, therapeutic targets, and therapeutics implicated in aging, longevity, and disease.

 

About Insilico Medicine

Insilico Medicine is a pioneering global biotechnology company dedicated to integrating artificial intelligence and automation technologies to accelerate drug discovery, drive innovation in the life sciences, and extend healthy longevity to people on the planet. The company was listed on the Main Board of the Hong Kong Stock Exchange on December 30, 2025, under the stock code 03696.HK.

By integrating AI and automation technologies and deep in-house drug discovery capabilities, Insilico is delivering innovative drug solutions for unmet needs including fibrosis, oncology, immunology, pain, and obesity and metabolic disorders. Additionally, Insilico extends the reach of Pharma. AI across diverse industries, such as advanced materials, agriculture, nutritional products and veterinary medicine.

For more information, please visit www.insilico.com

ON THE GROUND

U.S. Midterms 2026: 

Georgians on the front lines of the people’s war against data centres

Artificial intelligence emerged as a major issue on the 2026 US midterm campaign trail, with opposition to data centres drawing supporters across partisan divides united over ecological worries, increased utility bills and business interests outweighing citizen concerns. Georgia, the top US market for data centre construction, has been on the front line of the struggle.

Issued on: 02/11/2026
FRANCE24
By: Leela JACINTO

File photo of a Meta data centre near Social Circle, Georgia, USA, taken September 5, 2026. © Mike Stewart, AP


Quilliam Barber surveyed his backyard on a beautiful fall morning – the towering trees shading the southern sun, early autumn leaves dotting the green lawn, a tuneful bird working the quiet of this rural Georgia district to its advantage – and let out a heavy, forlorn sigh.

“I feel terrible. It's right there,” said the 67-year-old retired truck driver, gesturing towards the sprawling open land behind his backyard. “It’s right up under my arm, you know. And I can do nothing about it.”

Quilliam Barber and his wife, Wilhemina Barber, take a break in their backyard in Coweta County in the southern US state of Georgia. © Leela Jacinto, FRANCE 24

The Barbers’ property abuts a massive, 800-acre tract earmarked for a hyperscale data centre in Georgia’s Coweta County in the Middle Chattahoochee River Basin, part of the wider Apalachicola Basin in southeastern USA.

His new neighbour, the Project Sail data centre, made headlines in early 2025, when the eye-popping scale of the plan came to light. The $17 billion venture, estimated to require roughly 900 megawatts of electricity and 9 million gallons of water, includes 13 buildings, each covering 378,000 square feet, taking the total land area to more than 4.9 million square feet, according to county filings.


Standing next to her husband, Wilhemina Barber knows that with such a mega project coming up next door, things are about to change – dramatically. “It’s very quiet and peaceful out here. We love it, I love it, he loves it. We’ve got a pretty clear feeling it's going to play an important role if they do put it right in our back door," she said.

Around two miles down the road, Ron Bockrath, picked up a poster from a pile on his kitchen buffet table. “Protect Coweta: No Data Center Here” read the yard sign in bold red.

Ron Bockrath puts up a resistance at home in Newnan, Coweta County, Georgia. © Leela Jacinto, FRANCE 24

Bockrath’s opposition to the data centre has kept him busy over the past few months. From organising collectives and filing lawsuits, to collecting petition signatures, his efforts are part of a multi-pronged community campaign to stop the Sail project.

Across the US, in cities and counties rich or poor, rural, urban or semi-urban, citizens are raging against the machine. Artificial Intelligence (AI), and the mega centres required to power it, emerged as one of the major campaign issues ahead of the November 3 midterm elections, popping up in debates and town hall gatherings, leaving many candidates scrambling to articulate a cogent position and their constituents dissatisfied. The public sentiment has been at odds with US President Donald Trump’s promotion of AI as “Superior Intelligence” (SI).

Community-led initiatives have targeted AI-driven surveillance systems, including cameras misused by law enforcement officials.

The backlash has at times been spearheaded by software engineers and geeks capable of hacking systems to identify guardrail weaknesses, and the gap between spin and reality. But with data centres, the mega infrastructure required to power AI, the pushback has been sweeping, drawing together Democrat and Republican supporters in a rare display of bipartisanship in a bitterly divided country.

That’s because they represent a confluence of concerns: from rising electricity bills, to environmental hazards affecting local communities and impacting global warming, to a public dissatisfaction over millionaires, billionaires, tech mandarins and financiers profiting from the new technology amid a cost-of-living crisis.

Georgia, a state that defined itself as a symbol of the “New South” after the American civil war, has kept up with the times. A technological boom, centred around Atlanta – “The capital of the New South” – has seen the area overtake Northern Virginia as the country’s top market for data centre construction.

It has left Georgians on the front lines of pitched, local battles in the people’s war against data centres.
A revamp from ‘rural conservation' to industrial district

At the entrance to the Bockrath property in Newnan, Western Coweta County, another yard sign proclaims the mission that has occupied the 83-year-old retired research scientist for months.

“NO Industrial Zoning. NO Data Center,” reads the sign, referring to a rezoning of this rural area into an industrial zone.

A yard sign outside the Bockrath residence in Newnan protests the rezoning of Coweta County from a rural to industrial area. © Leela Jacinto, FRANCE 24

In April, county commissioners in this traditionally Republican stronghold approved a zone code revamp from “rural conservation district” to industrial area by a narrow 3-2 vote.

Bockrath attended the meeting in a room packed with county residents, activists and journalists – and was dismayed by the proceedings.

“People were supposed to be allowed to speak – for or against. That's good if you've got a population that's somewhat equally divided, then it makes sense to give equal time to both sides. However, in this case, the local population, not just the Western Coweta County, but the entire county, wanted nothing to do with these data centers and the way they were being forced upon us. But they gave equal time to each side to speak,” he explained.

Project Sail is just one of at least five planned data centres in the county. Residents and collectives opposed to the various projects come together for countywide decision meetings. They also keep an eye on community initiatives across Coweta County, located around 50 miles southeast of Atlanta and home to around 146,158 people.

“Twenty minutes was allowed for opposition, which isn't really very long, considering the magnitude of this project. That may be legally okay, but morally, it’s not okay. I got there almost two hours before the meeting. The 20 minutes were up before I was allowed to speak. They’re supposed to give great credence to people living in the area, but they were not allowed to express their opinion,” he said.

The 2-3 vote was hailed by the commissioners who approved the rezoning. “This is a no-brainer for the county,” Commissioner Alphonso Smith told the Atlanta Journal-Constitution. “I will stake my political career on it.”

A representative of the project’s developer, Prologis, a Fortune 500 company, said the rezoning would be transformative for the area and would generate tax revenues, the local newspaper reported. “The question isn’t if this site will be developed. It’s just how,” JC Witt, Prologis senior vice president of data centre investments told the commission meeting.

‘Miles of backup traffic’ for a signature drive

The rezoning vote sparked multiple initiatives aimed at reversing the decision.

Bockrath, along with some of his neighbours, including the Barbers, promptly filed two lawsuits in a Coweta County court, challenging the rezoning. The cases, filed in May, are currently making their way through the legal system.

Meanwhile a grassroots collective, Citizens for Rural Coweta, launched a campaign for a “home rule petition”, which under Georgia's Constitution allows residents to petition the county's court to amend or repeal resolutions. In counties with more than 50,000 residents, signatures from at least 10% of registered voters are required.

The signature drive saw Bockrath hit the road, knocking on doors to inform neighbours about the initiative. Local businesses across the county opened their doors for petition signings. Every week, venues and schedules were posted on the group’s social media accounts, followed by weekly updates.

The response was overwhelming, said Bockrath. “There were miles, literally miles of backup traffic to come in and sign these things,” he recounted with a smile.

Finally, on Friday, October 16, Citizens for Rural Coweta volunteers submitted boxes with more than 14,800 signatures to a county court. The 10% mark had been reached. Under Georgia's Constitution, the probate judge has 60 days after a petition is filed to determine whether it’s valid. If it is, the judge must call an election to be held between 60 and 90 days after the filing.
‘We need to speak the truth’

Community-led campaigns against data centres are taking up time and considerable mobilisation effort from citizens across the US.

Bockrath admits it can be discouraging at times but he believes he has no choice. “Somebody has to do it,” he maintained. “We can't just give up. I don't think we want to see our country, or the world for that matter, get that messed up.”

Volunteers have been drawn to the movement for a variety of reasons and causes, from the hyperlocal to global. For Bockrath, the motivation is primarily scientific.

The retired researcher is concerned about the massive amounts of water required to cool systems in mega data centres. In Georgia’s Fayette County, for instance, many households face low water pressure complications, while some Newton County residents have been unable to drink the murky, brown water trickling out of their taps.

Energy consumption and the impact on global warming is an issue that particularly worries Bockrath. “If you keep using more of our electricity, generating more heat, climate change is going to accelerate,” he said urgently before launching into a mini-lecture on global temperature percentages and melting polar ice caps.

Underlying all these issues, Bockrath notes, is the spinning of truth. “In studying science, one of the things you learn is that you don't need to necessarily agree with your advisor, or chairman. What you're trying to do is find the truth,” he explained. “I'm not sure if that's the case always in things like marketing, in sales. There, I think, the idea is to sell your product or your idea, truth may not be as important. But I think it is. And I think we need to speak the truth.”

Finding ‘another place to go’

Georgia’s data centre boom has been driven in large part by tax breaks and low electricity rates granted to many projects under outgoing Republican Governor Brian Kemp, who has held the state’s top post since 2019.

But while companies get cheap rates, the cost of electricity for households has been rising exponentially. New data centres are driving up utility bills in at least 13 states, according to the Institute for Energy Economics and Financial Analysis. A Bloomberg analysis last year found residents living near data centres were paying as much as 267% more a month for energy than five years ago.

The increase in utility bills is also adding to the cost-of-living crisis fueled by rising gas prices due to the Iran war. Coweta residents such as Bockrath are not persuaded by the arguments made by developers, and accepted by some commissioners, that data centres bring in tax revenues.

“At the state level, we've got Governor Kemp, who's been very, very pro-business. I think he's been so pro-business that he's forgetting about the citizens,” he said. “You can bring in tax revenue from a data center, but if the citizens are being made to bear that cost in terms of their electrical bills, how is that helping them? It's not. It's hurting them.”

It’s also hurting across ideological lines, making the resistance to AI, and data centres in particular, a rare bipartisan issue as the US prepares to vote in the midterms.

While Coweta County, like most rural areas outside Atlanta, has long been a Republican stronghold, Bockrath notes that volunteers never discuss their partisan inclinations. The octogenarian grandfather describes himself as an independent, voting "both ways" in the past, depending on candidate platforms and performance. Although he declines to divulge how he plans to vote on November 3, Bockrath expresses disappointment with Republicans who have failed to challenge some of Trump's policies. "The Republicans didn't say, hey, that's not right, which they should have," is all he's willing to say.

The anti-data centre movement, he explains, straddles partisan divides. Taking the example of the recent signature drive, Bockrath guesses that there had to be a mix of supporters of both parties. “To pick up almost 15,000 registered voters to sign a petition here, many of them had to have been Republican. Also, a number of them, I'm sure, are Democrat. It's not that much of a political issue. It shouldn't be. This is a humanity issue,” he said.

For Barber, it’s also a matter of being forced to lose his past and links to the land.

“I might have to relocate if I can. I feel terrible. I love this area. I grew up in this area. I raised my family here. I had kids, my grandkids, they all played, we played, in the woods out there,” he said. “But if they made me an offer, then yes, I would relocate.”

His wife listens silently, her eyes fixed on the grass. “I'd be sad. I'd be very sad because it's just like starting all over again,” she revealed. “But if push comes to shove, we've discussed it, if they offer us a price and if it's something that we can live with, we'll take it and we'll find another place to go.”