Neowin News Feed for: Hallucinations

Deloitte agrees to refund Australian government after AI hallucinations found in report

Usama Jawad — Mon, 06 Oct 2025 22:14:01 +0000

Accounting firm Deloitte has agreed to partially refund the Australian government after it was essentially proven that its report contains instances of content generated by AI. Read more...

OpenAI bakes "CriticGPT" model to find mistakes in ChatGPT's answers

Aditya Tiwari — Thu, 27 Jun 2024 19:40:01 +0000

OpenAI has trained a new model called CriticGPT to catch LLM bugs. As the name says, it's meant to analyze ChatGPT's code outputs during the training process and provide critiques for human trainers. Read more...

Microsoft is making a new tool to block and rewrite ungrounded info for better AI responses

John Callaham — Fri, 21 Jun 2024 13:48:01 +0000

Microsoft says it is trying to cut down on hallucinations in responses from generative AI like Copilot with an upcoming tool that's designed to both block and rewrite ungrounded info. Read more...

Bug or a feature? AI “dream machines” actually don't have hallucination problem, expert says

Martin Hodás — Wed, 13 Dec 2023 21:08:01 +0000

AI-based chatbots often come up with factually incorrect answers, making themselves an easy target for criticism. But are LLMs really broken? AI expert argues hallucinating is their greatest feature. Read more...

Anthropic launches Claude Instant 1.2 which hallucinates less than other models

Paul Hill — Thu, 10 Aug 2023 20:08:01 +0000

Anthropic has announced the availability of Claude Instant 1.2 through an API. Interestingly, benchmarks show it to be the safest of all the models so far with fewer hallucinations. Read more...

Call of Duty anti-cheat update makes cheaters "hallucinate" decoy opponents

Pulasthi Ariyasinghe — Fri, 30 Jun 2023 14:16:01 +0000

Call of Duty's Ricochet Anti-Cheat team is now making cheaters see hallucinations that mimic a regular player but are only visible to the cheating software, flagging them in the system. Read more...

OpenAI develops LLM that uses a chain of thought like humans

Paul Hill — Wed, 31 May 2023 21:50:01 +0000

OpenAI is currently researching process supervision, a method for reducing the number of hallucinations produced by LLMs. It has had success with maths datasets but needs to be tested generally. Read more...