TechCrunch Startup News
OpenAI’s new reasoning AI models hallucinate more
Apr 22, 2025
OpenAI has launched new AI models, o3 and o4-mini, touted as state-of-the-art. However, these models surprisingly exhibit higher rates of hallucination than their predecessors. This increase raises worries about their reliability in professional settings. The discussion delves into the complexities of hallucinations as a persistent challenge in AI development, supported by expert insights and research findings.
05:21
AI Summary
AI Chapters
Episode notes
Podcast summary created with Snipd AI
Quick takeaways
- OpenAI's new O3 and O4 Mini models exhibit a significant increase in hallucinations compared to earlier models, raising reliability concerns in critical applications.
- Integrating web search capabilities into AI reasoning models shows promise in enhancing accuracy and reducing hallucinations by providing real-time information retrieval.
Deep dives
The Hallucination Challenge in AI Models
Recent AI models, specifically OpenAI's O3 and O4 Mini, struggle significantly with the issue of hallucinations, which involve generating false or misleading information. Despite advancements in coding and mathematical tasks, these new models have a higher hallucination rate compared to their predecessors, with O3 hallucinating in response to 33% of questions on internal benchmarks. This issue raises concerns about the reliability of AI in applications requiring accuracy, such as legal contracts, where factual errors can have serious consequences. Experts indicate that these hallucinations could stem from the reinforcement learning approaches used in these models, necessitating further research to address the growing problem.
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
