ITI vs MIT Technology Review

ITI: Affordability, Innovation Must Guide America’s Energy Future

Following congressional consideration of the High-Capacity Grid Act, the Load Forecasting Enhancement Act, the Affordable Innovation ...

ITI: Affordability, Innovation Must Guide America’s Energy Future

Following congressional consideration of the High-Capacity Grid Act, the Load Forecasting Enhancement Act, the Affordable Innovation ...

ITI Warns Broad FCC Supply Chain Rules Could Undercut Innovation

ITI urged the FCC to adopt a targeted, risk-based approach as it considers additional restrictions under its equipment authorization program.

AI Agents Fail at Self-Improving Open-Ended Research

A multi-institution research team led by Peter Kirgis and Sayash Kapoor at Princeton University examined whether AI agents can conduct open-ended scientific research. Their study (arXiv:2607.27191) finds that while agents can solve narrow technical tasks needed for AI research, they lack the judgment, creativity and qualitative decision-making required to produce original, conference-level research. The results suggest that recursive self-improvement and fully autonomous research by AI agents remain further off than some industry hype implies. The article notes prior evaluations focused on narrow, verifiable tasks and contrasts those with the demands of open-ended scientific inquiry; it also mentions that companies such as Anthropic and OpenAI remain confident in their systems' capabilities.

• A cross-institution research group led by Peter Kirgis and Sayash Kapoor at Princeton University published a study on AI agents (arXiv:2607.27191).
• The study concludes AI agents can solve narrow technical problems but cannot perform open-ended scientific research requiring judgment and creativity.
• Researchers found agents fall short of producing original results at the level expected by leading machine-learning conferences.

Is Research Over Because of AI? Eric Schmidt Weighs In

An MIT Technology Review commentary (published on t3n.de) examines whether AI breakthroughs like AlphaFold mean the end of traditional scientific discovery. The article recalls AlphaFold's ability to predict protein 3D structures, notes that Demis Hassabis and John Jumper of Google DeepMind were awarded a 2024 Nobel Prize for chemistry (as reported), and argues that AlphaFold sparked a wave of startups and large investments in AI for biology, chemistry, and materials. The piece contends that while AI changes how discoveries happen, it does not represent a literal end to scientific research and highlights data quality, scientific standards, and AI agents as areas of opportunity and challenge.

• Article published by MIT Technology Review Online on t3n.de on 2026-08-15.
• AlphaFold predicts three-dimensional protein structures by learning from thousands of experimentally measured protein shapes.
• The article states that Demis Hassabis and John Jumper of Google DeepMind received a 2024 Nobel Prize in Chemistry for AlphaFold.

Local AI Models Make Phishing Emails More Dangerous

Researchers at TU Berlin, Inria and Ruhr-University Bochum report that small, local AI language models substantially increase phishing effectiveness by producing highly personalized emails. In controlled tests, successful attack rates rose from about 4% to roughly 10% after AI-based personalization, and an email security system flagged only one of nearly 4,000 AI-generated messages. The study shows how lightweight models can automate tailored social engineering at scale. Interpol has warned that cybercriminals increasingly use AI to automate phishing campaigns, create deepfakes and fake identities, and evade traditional signature-based defenses. The article summarizing the findings was published on t3n by Wolfgang Stieler in August 2026.

• Study by researchers at TU Berlin, Inria and Ruhr-University Bochum tested AI-enhanced phishing emails.
• Small, local AI models increased successful attack rates from about 4% to roughly 10% in controlled tests.
• An email security system flagged only one of nearly 4,000 AI-generated phishing messages in the experiment.