OpenAI's macOS app now includes Computer History, a feature that monitors user activity to train AI models and suggest automations. The tracking is opt-in with granular controls.
ChatGPT's desktop application for macOS has introduced Computer History, which logs clicks, keystrokes, and other user actions to build an activity timeline. The system uses this data to train ChatGPT and Codex models, enabling AI to suggest automations and resume incomplete tasks.
Unlike many data collection features, Computer History defaults to opt-in rather than opt-out, giving users initial control over participation. Users can exclude specific applications and websites from tracking and manually delete individual entries for finer-grained privacy management.
The feature transforms user behavior into training data, allowing the AI to learn work patterns and workflows. This approach differs from traditional analytics by capturing granular interaction details rather than aggregate metrics.
OpenAI has positioned the feature as a convenience tool, though users concerned about data collection can disable it entirely or selectively restrict which apps are monitored.
A secondary market for AI API credits is developing, with traders buying and selling unused computational allocations. The practice raises questions about pricing efficiency and market dynamics in the AI services sector.
Renowned mathematicians Timothy Gowers and Peter Sarnak say large language models are skilled at combining existing methods but fail to generate genuinely novel mathematical ideas.
Anthropic has released system prompts functionality, allowing developers to customize Claude's behavior and responses for specific use cases without fine-tuning.
A Google-backed study found that preventing AI chatbots from claiming consciousness fundamentally alters their positions on animal rights, religion, and life satisfaction. The finding suggests constraints in one area cascade across an AI model's entire worldview.