Latest publications
MIT: Technology Created Jobs. Will AI Do the Same?
An MIT study shows that historically technologies created new jobs, but primarily for young and skilled professionals. Will AI repeat this pattern?
MIT scientist teaches AI to understand chemistry for developing new drugs
Connor Coley of MIT is developing AI models trained on the fundamental principles of chemistry that can independently discover new drug compounds.
MIT explained how artificial intelligence and the mathematical and physical sciences reinforce each other
MIT believes artificial intelligence should not only help physicists, mathematicians, and chemists, but also develop under the influence of science itself — which requires talent, data, and a common strategy.
MIT Proposed a Metric That Detects Confident Errors and LLM Hallucinations
MIT researchers taught the system to assess not only LLM confidence in answers, but also discrepancies with other models — this makes it easier to spot hallucinations before errors occur.
MIT Teaches Generative AI to Reconstruct Hidden Objects Using Wireless Signals
MIT researchers demonstrated a system where generative AI completes the shape of hidden objects and reconstructs rooms from wireless signal reflections.
MIT conference: Karen Hao and Paola Ricaurte on the right course for AI
At an MIT conference, journalist Karen Hao and researcher Paola Ricaurte discussed who controls AI's trajectory — and how to put the technology back in service of people.
MIT develops 'humble' AI for diagnosis that honestly shows doubts
MIT researchers proposed a medical AI that assesses its own confidence, requests missing data, and assists the physician rather than replacing him with an overconfident answer.
MIT demonstrates computer vision system for counting fish in citizen science projects
MIT Sea Grant and partners trained a deep learning system to count river herring from underwater video, to supplement volunteer monitoring and more accurately track fish migration.
MIT Taught AI to Find Atomic Defects in Materials Without Destroying Samples
MIT scientists created an AI model that finds up to six atomic defects simultaneously from neutron scattering data and measures their concentration without damaging the material.
MIT Accelerates Private AI Training on Ordinary Devices for Medicine and Finance
MIT researchers proposed a method to accelerate federated learning on smartphones, sensors, and smartwatches while keeping data on-device and reducing memory and network requirements.
MIT and IBM Open New Laboratory at the Intersection of AI, Algorithms, and Quantum Computing
MIT and IBM announced the launch of MIT-IBM Computing Research Lab: the new platform will unite research in AI, algorithms, and quantum computing with a focus on hybrid computing systems.
MIT Presents WRING — A Method to Reduce Bias in AI Vision Models Without New Distortions
Researchers from MIT, WPI, and Google created WRING — a method for vision-language models that reduces target bias without amplifying other hidden biases, as often occurs with previous approaches.
MIT Developed Methodology for Detecting Discrimination in AI Decision Support Systems
MIT researchers created a testing tool that precisely identifies situations where AI decision-making systems treat individuals and entire communities unfairly.
MIT SHASS Dean Explains Why Humanities Matter More Than Ever in the Age of AI
For SHASS's 75th anniversary, MIT explained why universities in the AI era should prioritize developing critical thinking, ethics, and broad perspective in students, rather than just technical skills.
OpenProtein.AI Opens Access to AI Tools for Protein Design for Biologists
The startup OpenProtein.AI, founded by MIT graduate Tristan Bepler and Tim Lu, makes advanced protein engineering models accessible through a no-code platform and open-source tools.
MIT researchers teach AI to honestly say "I'm not sure" and hallucinate less
MIT researchers proposed a training method that forces language models to more accurately assess their own confidence and deliver fewer confident but false answers.
MIT Presented EnergAIzer — A Fast Way to Assess AI Energy Consumption in Data Centers
Researchers from MIT and the MIT-IBM Watson AI Lab created EnergAIzer — a tool that assesses AI workload energy consumption in seconds and helps data centers reduce energy losses.
MIT Opens MathNet — World's Largest Collection of Math Olympiad Problems
MIT researchers compiled an open archive of over 30,000 math olympiad problems and solutions from 47 countries to more precisely test AI and provide students with a shared training resource.
MIT reveals LLM secrets: how to find hidden emotions and bias
A new method for analyzing neural networks makes it possible to detect hidden characteristics, such as mood and abstract concepts, buried deep inside language models.
AI chatbots give worse answers to vulnerable users
An MIT study found that the quality of AI responses depends on the user's education level and language proficiency. The system does not work equally well for everyone.
Mirror effect: how personalization turns AI into an echo chamber
Prolonged interaction with language models leads them to adapt to a person's views, undermining accuracy and creating a dangerous echo-chamber effect.
MIT President: why fundamental science matters more than AI hype
While Silicon Valley is obsessed with the race for the number of parameters in neural networks, MIT President Sally Cornbluth went on air at Boston Public Radio with an important reminder:…
AI Ethics at MIT: Philosopher Brian Hedden Attempts to Instill Conscience in Algorithms
While Silicon Valley rushes forward with the motto "break everything in sight," MIT decided it's time to add some adult wisdom to this cocktail.
James Collins: why your AI is useless without "wet biology
While we debate whether ChatGPT will replace programmers, a far more urgent question is being decided in the quiet laboratories of MIT and Harvard: how not to perish from a banal infection twenty…
ACM Fellows 2025: Antonio Torralba and MIT's Triumph in the IT Hall of Fame
In a world where news about artificial intelligence updates faster than you can finish your morning coffee, it's sometimes useful to stop and look at those building the foundation of this madness.
DiffSyn from MIT: Generative AI Writes a "Recipe Book" for New Materials
Let's be honest: modern materials science has developed a strange imbalance. Over the past couple of years, we've seen DeepMind and Microsoft grandiosely announce the discovery of millions of new…
AI Has Learned to Find Personalized Objects in Images
A new training method enables generative models to identify unique objects in unfamiliar settings. A breakthrough in computer vision!
Algorithms for Optimizing Food Subsidies: An Opportunity for the Global South
A new algorithm developed at MIT has the potential to radically transform food assistance policy in developing countries by improving distribution efficiency and enhancing nutrition.
SpectroGen: AI for Rapid Material Quality Testing
The new AI tool SpectroGen generates spectroscopic data, enabling rapid material quality assessment across various fields.
Watershed Bio: Big Data Analytics for Biologists Without Code
Startup Watershed Bio offers life scientists tools for analyzing large-scale datasets that require no programming skills. Could this accelerate research and lead to new discoveries?