AI chatbot hallucinations will be largely eliminated within about a year (by end of 2025), with outputs as accurate as web content.
Inaccurate ●●●●●
“I'm very convinced that we're going to largely eliminate hallucinations over the next year. And they're going to be as accurate as any content that you can find on the web”
- Said on
- 21 February 2024 · Nasscom Technology Leadership Forum 2024
- Source
- DeepMind's Mustafa Suleyman predicts end to AI hallucinations by next year · Communications Today
- Should come true
- 2025
- How sure they sounded
- I'm very convinced
- Categories
- AI
- Importance
- 3 of 5 · Reliability of LLM output is central to AI adoption across industries.
Outcome
Inaccurate
Hallucinations were not largely eliminated by the end of 2025: OpenAI's April 2025 o3/o4-mini system card reported PersonQA hallucination rates of 33% and 48%, higher than earlier models, and hallucination remained a widely acknowledged unsolved problem through 2025.
- https://simonwillison.net/2025/Apr/21/openai-o3-and-o4-mini-system-card/
- https://tech.yahoo.com/articles/openais-o3-o4-mini-hallucinate-173655834.html
History
- 2026-10-17 07:51:00 · created by claude
- 2026-10-17 07:51:00 · categorized by claude
- 2026-10-17 07:51:18 · scored by claude