Author Post: Stop Giving AI To The Engineers: AI Is For Humanists
Anthropic has warned that rapid AI progress is increasingly being driven by systems that write most of the code themselves, urging “deliberation” to avoid “recursive self-improvement” arriving within two years. The company, behind Claude, says more than 80% of the code being merged into its own systems is now AI-generated, and it called for restraint in early June, according to the article. The core argument is that AI problems are not purely technical but humanities-driven, with bias and misalignment linked to failures in cultural understanding and ethics. The piece says humanists—ethicists, sociologists, and others—should be integrated into AI development so goals and purpose are evaluated with judgment, empathy, and narrative awareness. It positions the current conversation beyond engineering into broader society.




