Israeli Researchers Develop AI Self-Teaching Method Without Human Data

Researchers at Tel Aviv University and Stanford University have developed a training method where two AI models generate challenges for each other and improve prediction without human-generated content or external datasets. Published on arXiv, the approach could change how future AI systems are trained.

• OpenAI Algorithms Reportedly Crack Navier-Stokes, Shaking Mathematics2026-10-23
• Stanford PhD student uses AI to design virus genetic blueprints2026-10-09
• People hate AI but can't stop using it2026-10-05
• Stanford criticized for AI-altering race, appearance of students in photo2026-09-28
• HomeBody: Long-horizon humanoid autonomy without policy learning2026-09-27