Artificial intelligence systems often perform impressively on standardized medical exams—but new research suggests these test scores may be misleading. A study published in JAMA Network Open indicates that large language models, or LLMs, might not ...

Stanford's 2026 AI Index: frontier models fail one in three attempts, lab transparency is declining, and benchmarks are saturating faster than they're replaced.

Cisco tested eight major open-weight artificial intelligence models and found multi-turn jailbreak attacks succeeded nearly 93% of the time. (Image: Shutterstock) Enterprise artificial intelligence deployments are running on models that fold nearly every ...

In practice, retrieval is a system with its own failure modes, its own latency budget and its own quality requirements.

Large language models typically perform so similarly that their differences can be measured by millimeters. But in some scenarios, these models are separated by miles. After a chance discovery that ChatGPT seemed more likely to return strange and unlikely ...

Poor models. They have to wear whatever other people tell them to, including treacherous footwear, and they have to walk through a room under the scrutiny of everyone present looking completely confident and physically perfect, and very, very serious.