# AI Now Learns by Watching, Not Just Following Instructions OpenAI and Anthropic announced features allowing AI systems to learn by demonstration. - **Source:** [Fast Company](https://www.fastcompany.com) - **Published:** 2026-08-19 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/bab18500-dff5-45cd-bbdc-534f66a5a751 ## A New Era in AI In the summer of 2023, OpenAI and Anthropic announced new features allowing AI systems to learn tasks by watching them performed once. This marks a significant departure from traditional instruction-based methods. OpenAI's Record Replay feature enables ChatGPT and Codex users to demonstrate workflows and turn them into reusable skills. ## Anthropic's Approach Anthropic introduced Record a Skill within Claude Cowork. This feature allows users to record their screen while performing a task and narrate their reasoning. Claude then converts this into a skill that can be executed again. This approach closely mirrors OpenAI's solution. ## Competition and Innovation Two competitors unveiled similar solutions within weeks of each other, signaling a new era in AI training. This is seen as an admission that instructions alone cannot take AI to the desired level. Learning by observation could enable AI to handle more complex tasks. ## Potential of Record Replay OpenAI's Record Replay feature offers significant convenience, especially for developers. Users can demonstrate workflows and automate them, saving time. This could increase efficiency in software development processes and reduce error rates. ## Advantages of Claude Cowork Anthropic's Record a Skill feature allows users to teach tasks more naturally. The combination of screen recording and voice narration helps AI better understand the logic behind tasks. This could be particularly useful for complex operations. ## The Future of AI Training Learning by observation could become a new standard in AI training. This approach enables AI to learn in a more human-like manner, allowing it to take on more complex and customized tasks. This could revolutionize the industry. ## Company Strategies The similar features introduced by OpenAI and Anthropic indicate a common need in AI training. Both companies are developing innovative solutions to make AI learning more effective and efficient. This competition could accelerate progress in the sector. ## The Future of AI Learning by observation features represent a significant step forward for AI. This approach enables AI to learn in a more human-like manner, allowing it to take on more complex and customized tasks. This could revolutionize the industry. ## Sources covering this story - https://www.fastcompany.com/91584431/prompts-arent-enough-why-two-ai-giants-are-evolving-their-approach-ai-prompts-technology — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._