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
If artificial intelligence (AI) systems shape decisions that affect people's lives, they should do so fairly. This should be a given considering that potential applications for AI include automated hiring systems, as well ...
May 12, 2026
0
2
Machine learning & AI
The nonprofit ARC Prize Foundation on May 1, 2026, released the results of a new benchmark: a test of an AI system's ability to solve a game. The results were striking—humans scored 100%, while the most advanced AI systems ...
May 12, 2026
0
3
Robotics
New artificial intelligence (AI) technologies currently in development could lead to an imminent future where machines no longer just process and analyze information, but can feel. In his new book "Perceptive Machines," futurologist ...
May 12, 2026
0
5
Consumer & Gadgets
From August 2026, an EU-wide AI regulation will come into force requiring the labeling of AI-generated content. However, a research team from the University of Bayreuth and Aalto University in Helsinki, Finland, has found ...
May 11, 2026
0
9
Hi Tech & Innovation
The coffee might be poured by a human hand, but behind the counter something far less traditional is calling the shots at an experimental cafe in Stockholm.
May 11, 2026
0
4
Machine learning & AI
In recent days, evolutionary biologist Richard Dawkins wrote an op-ed suggesting AI chatbot Claude may be conscious.
May 7, 2026
0
7
Machine learning & AI
NASA announced that it will launch the Nancy Grace Roman space telescope into orbit in September 2026, eight months ahead of schedule. The new space telescope is expected to deliver 20,000 terabytes of data to astronomers ...
May 5, 2026
0
5
Security
Cybercriminals have been struggling to adopt AI in their work, reports the first-of-its-kind study that analyzed a dataset of 100 million posts from underground cybercrime communities. The study is published on the arXiv ...
May 4, 2026
0
42
Consumer & Gadgets
Ultra-personalized artificial intelligence for assisted communication risks muting aspects of the user's identity and occasionally breaches privacy, according to a new study from a Cornell Tech doctoral student who trained ...
May 4, 2026
0
7
Business
As artificial intelligence (AI) becomes embedded in critical decisions about health, safety, finance, and governance, a key challenge is no longer whether people and AI will collaborate, but rather how to structure this collaboration ...
Apr 30, 2026
0
10
