Latest about Artificial Intelligence
Mathematicians devised novel problems to challenge advanced AIs' reasoning skills — and they failed almost every test
Current AI models struggle to solve research-level math problems.
Communicating with aliens one day could be much easier if we study the way AI agents speak with each other
Decoding emergent languages in AI can have many benefits.
Meet Evo, an AI model that can predict the effects of gene mutations with 'unparalleled accuracy'
Evo is a large language model that is not trained on words but on the genomes of millions of microbes. It can accurately predict the effects of mutations.
Large language models not fit for real-world use, scientists warn — even slight changes cause their world models to collapse
Large language model AIs might seem smart on a surface level but they struggle to actually understand the real world and model it accurately, a new study finds.
Future passenger planes could use AI to eliminate turbulence and maintain a smooth in-flight experience
Turbulence in airplanes could become a thing of the past with FALCON, a new AI system that helps vehicles learn how to adjust to turbulence within a matter of minutes.
New memory chip controlled by light and magnets could one day make AI computing less power-hungry
A new type of ultrafast memory uses optical signals and magnets to efficiently process and store data.
Computers normally can't see optical illusions — but a scientist combined AI with quantum mechanics to make it happen
A deep neural network was trained using quantum tunneling to mimic the human ability to view optical illusions.
AI 'can stunt the skills necessary for independent self-creation': Relying on algorithms could reshape your entire identity without you realizing
"If you constantly use an AI to find the music, career or political candidate you like, you might eventually forget how to do this yourself." Ethicist Muriel Leuenberger considers the personal impact of relying on AI.
'I'd never seen such an audacious attack on anonymity before': Clearview AI and the creepy tech that can identify you with a single picture
"Concerns about facial recognition had been building for decades. And now the nebulous bogeyman had finally found its form: a small company with mysterious founders and an unfathomably large database."
Scientists design new 'AGI benchmark' that tests whether any future AI model could cause 'catastrophic harm'
OpenAI scientists have designed MLE-bench — a compilation of 75 extremely difficult tests that can assess whether a future advanced AI agent is capable of modifying its own code and improving itself.
