AI News
A simple twist fooled AI—and revealed a dangerous flaw in medical ethics
Even the most powerful AI models, including ChatGPT, can make surprisingly basic errors when navigating ethical medical decisions, a new study reveals. Researchers tweaked familiar ethical dilemmas and discovered that AI often defaulted to intuitive but incorrect responses—sometimes ignoring updated facts. The findings raise serious concerns about using AI for high-stakes health decisions and underscore the need for human oversight, especially when ethical nuance or emotional intelligence is involved.
Concrete that lasts centuries and captures carbon? AI just made it possible
Imagine concrete that not only survives wildfires and extreme weather, but heals itself and absorbs carbon from the air. Scientists at USC have created an AI model called Allegro-FM that simulates billions of atoms at once, helping design futuristic materials like carbon-neutral concrete. This tech could transform cities by reducing emissions, extending building lifespans, and mimicking the ancient durability of Roman concrete—all thanks to a massive leap in AI-driven atomic modeling.
School of Architecture and Planning recognizes faculty with academic promotions in 2025
The faculty members’ work comprises multifaceted research and scholarship across a wide range of disciplines.
MIT Learn offers “a whole new front door to the Institute”
The AI-enabled platform serves as a hub for MIT’s lifelong learning opportunities.
A new way to edit or generate images
MIT researchers found that special kinds of neural networks, called encoders or “tokenizers,” can do much more than previously realized.
The unique, mathematical shortcuts language models use to predict dynamic scenarios
Language models follow changing situations using clever arithmetic, instead of sequential tracking. By controlling when these approaches are used, engineers could improve the systems’ capabilities.
Model predicts long-term effects of nuclear waste on underground disposal systems
The simulations matched results from an underground lab experiment in Switzerland, suggesting modeling could be used to validate the safety of nuclear disposal sites.
This “smart coach” helps LLMs switch between text and code
The CodeSteer system could boost large language models’ accuracy when solving complex problems, such as scheduling shipments in a supply chain.
Can AI really code? Study maps the roadblocks to autonomous software engineering
A team of researchers has mapped the challenges of AI in software development, and outlined a research agenda to move the field forward.
How to more efficiently study complex treatment interactions
A new approach for testing multiple treatment combinations at once could help scientists develop drugs for cancer or genetic disorders.
New AI system uncovers hidden cell subtypes, boosts precision medicine
CellLENS reveals hidden patterns in cell behavior within tissues, offering deeper insights into cell heterogeneity — vital for advancing cancer immunotherapy.
RealSense spins out of Intel to scale its stereoscopic imaging technology
RealSense is building vision tech that helps machines like robots and drones navigate and respond to a 3D world.
This magnetic breakthrough could make AI 10x more efficient
A groundbreaking step in AI hardware efficiency comes from Germany, where scientists have engineered a vast spin waveguide network that processes information with far less energy. These spin waves quantum ripples in magnetic materials offer a promising alternative to power-hungry electronics.
Changing the conversation in health care
The Language/AI Incubator, an MIT Human Insight Collaborative project, is investigating how AI can improve communications among patients and practitioners.
AI shapes autonomous underwater “gliders”
An AI pipeline developed by CSAIL researchers enables unique hydrodynamic designs for bodyboard-sized vehicles that glide underwater and could help scientists gather marine data.
Study could lead to LLMs that are better at complex reasoning
Researchers developed a way to make large language models more adaptable to challenging tasks like strategic planning or process optimization.
New postdoctoral fellowship program to accelerate innovation in health care
Launched with a gift from the Biswas Family Foundation, the Biswas Postdoctoral Fellowship Program will support postdocs in health and life sciences.
Exploring data and its influence on political behavior
In MIT's course 17.831 (Data and Politics), students are introduced to the power of analysis, visualization, and research-supported insight into political outcomes.
Robotic probe quickly measures key properties of new materials
Developed to analyze new semiconductors, the system could streamline the development of more powerful solar panels.
AI spots deadly heart risk most doctors can't see
An advanced Johns Hopkins AI model called MAARS combs through underused heart MRI scans and complete medical records to spot hidden scar patterns that signal sudden cardiac death, dramatically outperforming current dice-roll clinical guidelines and promising to save lives while sparing patients unnecessary defibrillators.
