The Archive

Mistral announces edge-optimized small models, positioning them as best-in-class for edge deployment.

Anthropic updates Responsible Scaling Policy with new safety commitments and evaluation standards for frontier model development.

ChatGPT fairness analysis reveals response variance based on user demographic signals in names.

MLE-bench introduces benchmark for evaluating AI agents on machine learning engineering tasks.

OpenAI details efforts to identify and disrupt deceptive and abusive uses of its models.

Anthropic outlines measures to prevent Claude misuse for U.S. election manipulation and disinformation.

Hearst's lifestyle and local news content integrated into OpenAI products via partnership.

Canvas feature enables side-by-side writing and code editing in ChatGPT interface.

OpenAI secures $4B credit facility from major banks including JPMorgan, Citi, Goldman Sachs.

OpenAI announces $6.6B funding round to advance AGI mission and scaling objectives.

OpenAI releases Realtime API enabling low-latency speech-to-speech applications for developers.

OpenAI extends fine-tuning API to support vision tasks with GPT-4o for custom image-text models.

OpenAI introduces prompt caching to reduce API costs for repeated input sequences.

OpenAI releases model distillation API to fine-tune smaller models using frontier model outputs.

Altera uses GPT-4o to create human-AI collaborative workflows.

OpenAI banned Rwanda-origin accounts using ChatGPT to generate partisan election commentary ahead of August 2024 polls.

OpenAI terminated accounts exploiting its API via Israel-based startup to spam gambling links on X.

OpenAI shut down A2Z multilingual influence operation targeting elections, Ukraine, and politics across platforms.

OpenAI banned US-origin account creating fake ChatGPT error message falsely claiming detection of Russian troll activity.

OpenAI shut down CyberAv3ngers accounts linked to Iran using AI to research industrial control systems and vulnerabilities.