Can researchers stop AI making up citations?
Sep 9, 2025 - Nature - OpenAI’s GPT-5 hallucinates less than previous models do, but cutting hallucination completely might prove impossible.
Apertus: a fully open, transparent, multilingual language model
Sep 2, 2025 -EPFL, ETH Zurich and the Swiss National Supercomputing Centre (CSCS) released Apertus today, Switzerland’s first large-scale, open, multilingual language model.
GPT-5: has AI just plateaued?
Aug 11, 2025 - GPT-5’s focus on routing queries to the best model could be an admission that LLMs can’t get much smarter.
OpenAI launches reasoning LLM that you can download and tweak
Aug 6, 2025 - One version of the gpt-oss large language model can run on a laptop, and performs nearly as well as the company’s most powerful models.
OpenAI’s o3 tops new AI league table for answering scientific questions
Jul 10, 2025 - o3 outperformed DeepSeek-R1 and Google’s Gemini-2.5-Pro at answering questions on natural sciences, health care, engineering, and humanities and social science.
The downside of a digital yes-man
Jul 7, 2025 - Sycophantic behavior in AI can mean trading accuracy for flattery.
Why we need mandatory safeguards for emotionally responsive AI
Jul 1, 2025 - Virtual chatbots that simulate conversations with famous actors or sci-fi characters can have real-world consequences.
Introducing Mu language model and how it enabled the agent in Windows Settings
Jun 23, 2025 - This model addresses scenarios that require inferring complex input-output relationships and has been designed to operate efficiently.
Musk's Grok retraining is harder than it looks
Jun 24, 2025 - AI researchers warn that efforts to politically shape LLMs are misguided and dangerous.
What Happens When People Don’t Understand How AI Works
Jun 6, 2025 - Despite what tech CEOs might say, large language models are not smart in any recognizably human sense of the word.
OpenAI's new reasoning AI models hallucinate more
Apr. 18, 2025 - OpenAI's reasoning AI models are getting better, but their hallucinating isn't, according to benchmark results.
Study: AI firms train models using more premium publisher data than they let on
Nov 5, 2024 - AI firms rely disproportionately on commercial publishers of news and media websites to train their LLMs