AI researchers from the Dfinity Foundation and Onicai have unveiled a Manifesto for Decentralized AI, aiming to ensure AI benefits users rather than large institutions. Onicai CEO Patrick Friedrich warns against AI centralization, which could lead to manipulation by powerful entities. The manifesto suggests using decentralized AI on open networks to enhance transparency and user control.

Onicai's CTO, Arjaan Buijk, highlights the need for niche AI models, as current focus on general-purpose models stifles innovation. Meanwhile, Anthropic CEO Dario Amodei predicts human-level AI could emerge as early as 2026, much sooner than previous estimates.