MIT Technology Review
AI news source. Articles are auto-selected and adapted by Hamidun News editors.
Latest publications
Open AI Models: Democratization vs. Cyber Threats
Anthropic has labeled GLM-5.3 a cybersecurity threat, but the debate over open models runs deeper: AI democratization versus safety. Nvidia and Anthropic sit on opposite sides.
AI Labs Are Buying Up Dead Startups' Data Archives. Defunct Companies' Records Are Worth Millions
Crypto startup founder Kilan Miskey got $25,000 for his company's archives. But that's just the start: SimpleClosure and Sunset have already sold 150+ data packages to AI companies. Google spent $10 million on Spirit Air
We Overestimate AI's Ability to Say 'No': Why Refusal Isn't the Answer
MIT Technology Review dissects the illusion of safety built on AI refusal. Companies train models to decline dangerous requests, but the mechanisms remain unreliable. The core problem: useful and harmful capabilities can
Stanford Researcher Behind AI-Designed Viruses to Explain to MIT TR How a Model Writes Genomes
In 2025, Stanford PhD student Samuel King got a generative model to design virus genomes. MIT Technology Review will discuss the work with him on October 16.
Anthropic links Z.ai's new open model to a risk of large-scale cyberattacks
Anthropic's report says a powerful open model from China's Z.ai poses a serious cyberattack threat. MIT Technology Review asks whether the benefits of open weights outweigh the risks.
MIT Technology Review Warns That We Trust AI Model Refusals Too Much
Chatbots are trained to refuse a huge range of requests, and the industry treats that as protection. MIT Technology Review argues that faith in AI's ability to say "no" is overstated.
Flytrex Is Hiring RPICs: What an Air Traffic Controller for Delivery Drones Actually Does
Trevor Wisniewski heard about drones delivering pizza in his Texas neighborhood and took a job as a remote pilot at Flytrex. Here is what an RPIC does and why the role exists.
Why Human Control Over Combat AI Is an Illusion, and What It Means for the Pentagon
The Anthropic-Pentagon dispute reveals that a human in the loop of combat AI can only formally approve decisions without understanding the model's hidden logic and true intentions.
Pentagon prepares AI training on classified data, as new reactors change waste management approach
Pentagon wants to train military AI versions on classified data in secure circuits, and a new generation of nuclear reactors poses new waste management questions for the industry.
Pentagon Preparing Closed Circuits for Training AI Models on Classified Data
The U.S. Department of Defense is discussing secure facilities where developers can fine-tune military versions of generative models on classified information for more accurate military analysis.
OpenAI moves into US military systems, while xAI faces lawsuit over Grok and teen imagery
OpenAI is preparing its models for classified Pentagon systems, while xAI faces a lawsuit over Grok, which plaintiffs claim was used to create illegal sexualized images of minors.
OpenAI could enter U.S. military systems against Iran — from target selection to drone defense
After the Pentagon deal, OpenAI technologies could appear in the conflict around Iran — from strike prioritization to counter-drone defense and GenAI.mil.
Pentagon and Anthropic dispute Claude amid plans to use AI for target selection
The US Department of Defense is considering the use of generative AI for target prioritization while also pressuring Anthropic to ease Claude's restrictions for military use.
Pentagon may use AI chatbots to prioritize targets in military strikes
The US military could use generative AI to rank target lists and recommend the order of strikes, but the final decision, an official said, should remain with humans.
Deepfake porn as a weapon of harassment: how AI creates a new form of violence
The story of a woman who found synthetic porn featuring her face via facial recognition reveals the scale of the crisis: freely available tools can create compromising video content in minutes.
From technology to customers: how McKinsey suggests rethinking digital investments
A McKinsey study found that companies capture less than a third of the value of digital investments because of the wrong approach — they build technology instead of solving customer problems.
Anduril and Meta develop AR glasses for military drone control
Announcement from Anduril shows how future combat systems are evolving: a commander looks at a wall in a room, blinks — and a combat drone receives an order in a fraction of a second.
From Big Pharma to Nature: How a Merck Chemist Became a Natural Drug Designer
Tim Chernack spent two decades developing precision therapy at Big Pharma, but in 2018 chose a different path: bio-inspired design of molecules from nature.
David Sinclair to Test Anti-Aging Drugs in XPrize Competition
Harvard gerontologist David Sinclair plans to test the effectiveness of anti-aging drugs in the international innovation competition XPrize.
Heat disrupts brain function: scientists are beginning to understand why
Europe's heat wave has broken records, but its effects go beyond heatstroke — scientists are documenting cognitive and mental health effects of high temperatures.
Europe's heat shuts down power plants as IBM challenges Moore's law
Europe is enduring record heat: power grids are under strain, and several power plants have been forced to shut down. At the same time, IBM unveiled a next-generation chip aimed at extending Moore's law.
Extreme heat in Europe: how record temperatures are straining power grids
Extreme heat has swept across Europe, setting temperature records: schools are closing, deaths are being reported, and power grids are operating at the limit in the height of summer.
Heat in Europe overheats rivers and halts power plants — France sets a record unseen since 1947
Europe is facing its worst heat in decades: France broke the 1947 temperature record, some power plants were idled, and energy demand is hitting historic highs.
Stripe, Anthropic and OpenAI fund efforts to combat respiratory infections
Payment company Stripe and leading AI labs Anthropic and OpenAI announced support for a new scientific project to prevent the common cold and other respiratory infections.
Casey Harrell, who has ALS, becomes the first "power user" of a speech neuroimplant
Casey Harrell, who has amyotrophic lateral sclerosis, is the first patient described as a "power user" of a brain-computer interface. He uses his brain implant every day at home.
Data Centers Connect to the Grid Faster Through Flexible Power Consumption Agreements
A new flex-connection strategy lets data centers skip a multi-year queue: they agree to cut consumption at peak hours in exchange for faster access to the grid.
Claude Science from Anthropic: AI tool for scientific research based on Opus 4.5
On June 30, 2026, Anthropic introduced Claude Science—a flagship product for science that works like Claude Code but for research and biopharmaceuticals.
AI Agents as Office 'Colleagues': Why This Is a Flawed Metaphor — MIT Technology Review
MIT Technology Review explains why companies are wrong to call AI agents 'colleagues' and 'subordinates' at work, and why this trendy metaphor is dangerous in practice for employees.
AI in Agriculture Stalls Not Because of Technology, But Because of Data — MIT Technology Review
MIT Technology Review warns: the agricultural sector is ready to adopt artificial intelligence, but farm data is not yet ready for it and requires advance investment.
How AI helps small e-commerce sellers decide what to produce
AI analyzes demand and purchase history, helping small online store owners make decisions about product assortment more accurately and quickly.