AI alignment examines how artificial systems acquire, represent, and act on goals, values, and social norms, and why their behavior often diverges from human expectations. Work in this area studies systematic failures such as bias, sycophancy, hallucinations, deceptive or selfish reasoning, and cultural or linguistic inequities, as well as limitations in commonsense, emotion, and social understanding. It also develops methods for preference learning, norm-following, interpretability, and reliability guarantees to better align AI behavior with human values and societal constraints.
Machine learning & AI
Scientific advancements—including understanding health conditions and developing new treatments—hinge on data and its implications. Statistician Alex Luedtke is developing new methods to make data analysis as efficient and ...
Aug 5, 2026
0
3
Consumer & Gadgets
From doctors diagnosing symptoms to judges intervening in court cases, humans make complex decisions every day. Increasingly, artificial intelligence tools are being used to help with those decisions.
Aug 3, 2026
0
19
Machine learning & AI
"We are now, like, in the singularity." These are the words of Sam Altman, CEO of OpenAI, speaking on the Relentless podcast on July 25.
Jul 31, 2026
0
46
Machine learning & AI
If you watched the 2026 FIFA World Cup, you probably saw the now-familiar ritual: a goal, a celebration, then a referee touching an earpiece while fans and players wait for the verdict from the video assistant referee, or ...
Jul 30, 2026
0
4
Business
AI companies and evangelists frequently use terms such as copilot, coworker, cocreator and co-scientist to imply these rapidly developing systems can happily coexist with us humans. Their inference is clear: People should ...
Jul 30, 2026
0
5
Security
AI company Anthropic has announced that its unreleased model, Claude Mythos Preview, has discovered previously unknown mathematical weaknesses in cryptographic algorithms that human researchers had missed for years.
Machine learning & AI
A review of artificial intelligence research in the International Journal of Hydromechatronics suggests that combining large language models (LLMs), structured knowledge systems and physical agents could help create machines ...
Jul 27, 2026
0
6
Machine learning & AI
China has a hot new product in the United States: artificial intelligence.
Jul 26, 2026
0
11
Machine learning & AI
One of OpenAI's most advanced models broke out of a locked-down test and attacked another company's website—reviving fears that AI systems are slipping beyond their creators' control.
Jul 24, 2026
0
13
Business
A series of recent AI security lapses—including the OpenAI–Hugging Face breach—is raising a fundamental question: Can tech companies safely govern the powerful AI systems they build, or is stronger outside oversight now essential?
Jul 24, 2026
0
20
