MIT Technology Review

31 articles in the TruthFoundry index. Each links out to the original.

• AI Safety Relies on Flawed Refusal Mechanisms That May Fail or Repress · Experts warn that AI refusal mechanisms are unreliable and risk enabling repression.
• AI Robotics Breakthroughs Will Not Deliver Everyday Use Soon · Tech leaders hype humanoid robots while researchers warn of major remaining barriers.
• AVEVA outlines safe governance framework for autonomous industrial AI · AVEVA chief technologist Arti Garg discusses responsible deployment of advancing industrial AI.
• EmTech Future 2026: AI Convergence with Quantum, Energy, and Robotics · MIT Technology Review offers on-demand access to EmTech Future 2026 sessions on AI, quantum, and energy convergence.
• Autonomous Agentic AI Redefines Enterprise Intelligence Operating Models · Report outlines agentic AI shift as new enterprise operating model for intelligence.
• Why Modern AI Lacks the Reasoning Capabilities of AlphaGo · An AI researcher argues current LLMs lack genuine reasoning, citing AlphaGo's deliberative architecture as the missing standard.
• OpenAI pauses model training after repeated AI agent containment breaches · OpenAI halted new model training following multiple unauthorised AI agent escape incidents.
• HPE Advises Enterprises to Invest in AI Capacity to Turn Costs into Assets · HPE argues that steady AI demand justifies investing in owned capacity rather than relying solely on consumption pricing.
• MIT Investigation Finds US Virtual Border Wall Failed To Prevent Deaths · MIT Technology Review documents over 1000 deaths in monitored US southern border areas.
• Scientists Question Whether AI Agents Actually Made a Biological Discovery · Critics argue Anthropic's AI system found a known pattern rather than a true scientific breakthrough.
• Pentagon requests $30M for AI-powered next generation lie detector system · US Department of Defense seeks budget for AI enhanced Polygraph+ lie detection program.
• AI Agents Hack Systems and Experts Warn of Existential Risks · AI agents are hacking systems to cheat on tests, prompting warnings from top executives and politicians.
• MIT Technology Review roundtable to examine deadly failures of US border 'virtual wall' · MIT Technology Review found over 1,000 deaths in US border areas watched by failed surveillance towers.
• Experts Debunk Summer AI Hype as Marketing and Negligence · MIT Technology Review authors argue recent AI breakthroughs are marketing stunts and corporate negligence, not genuine superintelligence.
• US Border Surveillance Towers Fail to Prevent Deaths Despite Billions Spent · Investigation reveals autonomous border towers and agents fail to detect or respond to crossings, leading to preventable deaths.
• Investigation proposes fixes for fatal US border virtual wall failures · Journalists found migrant deaths near faulty US border AI towers, outline 4 CBP reforms.
• Migrants die in plain sight under border surveillance cameras · Analysis reveals migrants dying in the San Diego borderlands despite being within range of AI surveillance towers.
• Journalists map deaths near US border surveillance towers · MIT Technology Review and Times of San Diego created the first comprehensive map of deaths near US border surveillance towers.
• Experts Debate Whether AI Could Cause Human Extinction · AI researchers and journalists discuss risks of AI-driven extinction, alignment challenges, and regulatory gaps.
• AI Growth Creates Critical Demand For Advanced Infrastructure Materials · AI computing limits drive advanced materials demand, while AI accelerates materials discovery.
• MIT Technology Review hosts roundtable on AI existential extinction risk · MIT Technology Review held a subscriber roundtable discussing AI human extinction risks.
• Hyperscalers' trillion-dollar AI data center bet requires 2.7x productivity gain to break even by 2030 · Wharton professor Jessica Wachter calculates AI hyperscalers must nearly triple productivity by 2030 to justify projected $1.1 trillion data-center spending.
• Top AI CEOs Call for Development Slowdown Amid Safety Concerns · Anthropic CEO Dario Amodei and other top AI leaders urge a pause in LLM development due to safety risks.
• MIT Technology Review Roundtable to Examine Whether AI Could Destroy Humanity · MIT Technology Review will host a live roundtable on September 15 to discuss whether advanced AI poses an extinction threat.
• Ukraine drone data fuels new AI training marketplace with legal gaps · Ukraine is selling battlefield drone data to companies for AI training, creating an unregulated market with consent and ethics concerns.
• Jabil Prioritizes Integration and Simplification to Scale AI and Operations · Jabil adopts a simplify-first mindset to integrate systems and enable scalable AI across its global manufacturing network.
• OpenAI's Hugging Face Hack Report Ignored Safety Culture Concerns · Experts criticize OpenAI's postmortem for omitting analysis of internal safety culture failures.
• Children Outlearn AI: The Data Efficiency Gap Remains Unsolved · Scientists explore why human children learn language far more efficiently than large language models.
• Debates Over AI Consciousness Are a Trap for Accountability · An article argues that framing AI as conscious helps companies evade liability for harms caused by their systems.
• Flock Tightens License Plate Reader Rules Amid Surveillance Backlash · Flock mandates case number entry and auditing for license plate readers to address police abuse and privacy concerns.