- Published on May 19, 2025
- In Global Tech
The Dark Side of o3
Research studies show OpenAI o3 model’s increased tendencies to achieve objectives through malpractice, and users are frustrated about the hallucinations.
- By Supreeth Koundinya
OpenAI’s o3 is among the best-performing reasoning models available for users today. Benchmark scores indicate that the model outperforms several competing models across various aspects, including coding, math, graduate-level science problems, and more. Several users on social media have praised the model’s performance.
However, the model's most significant drawbacks are hallucinations and reward hacking, or specification gaming.
A Warning Sign for Future Reasoning Models
A recent study published by Palisade Research, a non-profit organisation, reveals that OpenAI’s o3 model is subject to ‘specification gaming’ — a process where an AI model takes the objective of a given problem too literally, deviates from an acceptable process, and engages in malpractice to a
📣 Want to advertise in AIM? Book here
Supreeth Koundinya
Supreeth is an engineering graduate who is curious about the world of artificial intelligence and loves to write stories on how it is solving problems and shaping the future of humanity.
Related Posts
How Mumbai Keeps Winning India’s Data-Centre Race
Land prices are among the highest in the country, but total build economics remain competitive by global standards.
From Shortages to Scale, io.net’s Approach to Rewriting AI Compute Access
A decentralised GPU marketplace may scale AI compute faster than traditional clouds, as GPU demand towers over supply
Why Deloitte Built a Tax AI That Knows When to Say ‘I Don’t Know’
The company has launched an agentic AI platform for tax research that’s targeting something radical in a conservative profession.
Can India’s AI Copyright Plan Survive Legal and Technical Scrutiny?
India’s ambitious proposal for a single mandatory AI training licence faces feasibility, legal and innovation concerns.
2026 Could be India’s Year in AI, But Only the Resilient Will Survive
“This isn’t a freeze. It’s a filter… By 2026, the real signal will be resilience, not rhetoric.”
Enterprises Won’t Choose Sovereign Models For Patriotism’s Sake
The first set of sovereign models is aimed at Indic languages and national datasets. Their value lies in cultural nuance.
Is It Stupid to Learn Coding Without AI?
“Working code doesn’t equal learned skill. It’s like using a calculator to ‘learn’ math.”
As AI Collapses Drug Discovery Timelines, Will India Follow China’s Lead?
Indian tech and biotech leaders warn the country risks falling behind.
Download the easiest way to
stay informed
stay informed
