Reports & Analysis: Deep Analysis

A critical analysis of sycophancy in AI and the changes ahead.

Explore new findings on how misalignment in language models occurs and methods to reverse it.

Explore the worst-case risks of open weight LLMs with insights on malicious fine-tuning in biology and cybersecurity.

Explore OpenAI's strategies to combat AI misuse in our in-depth October 2025 report.

Explore the capabilities and safety evaluations of the gpt-oss-safeguard models.

Explore our report on advanced AI adoption disparities and initiatives for productivity gains.

In-depth report on GPT-5 adoption trends and its application in various industries.

ANALYSIS