Human-level AI could be here as early as 2026: Anthropic CEO
Anthropic CEO Dario Amodei predicts that human-level artificial intelligence could be achieved as soon as 2026 or 2027, given the current rate of advances in AI technology. In a conversation with podcaster Lex Fridman, Amodei noted the progression of AI capabilities, likening them to educational levels: with AI currently at 'PhD level' compared to previous years. He acknowledged that factors such as data limitations or geopolitical issues could hinder this progress. Amodei highlighted the dual nature of powerful AI, capable of both positive and negative impacts, emphasizing the need for responsible development. He also commented on competition in the AI space, aiming for a 'race to the top' approach. While asked about the timeline for Anthropic's chatbot Claude Opus 3.5, he refrained from providing specific details.
Source 🔗