Anthropic CEO Dario Amodei calls to slow down AI development
Anthropic boss Dario Amodei outlines a three-point plan to slow AI model development, as industry leaders debate safety standards.
Stock photo for illustration only, not from the actual event
- Anthropic CEO Dario Amodei calls for a slower pace in developing artificial intelligence models.
- His proposed three-point plan includes independent monitoring, industry regulation, and global rules.
- OpenAI CEO Sam Altman and Elon Musk share contrasting views on AI capabilities and safety.
- Critics suggest the proposal may serve to consolidate control over advanced AI technology.
The head of leading artificial intelligence company Anthropic, Dario Amodei, has called for a deceleration in the development pace of AI models alongside stricter monitoring mechanisms.
Writing in an essay published on Saturday, Amodei stated that while developing artificial intelligence is inevitable, the associated risks are serious and both corporations and governments must be given adequate time to address them.
He introduced a structured three-point framework to guide future AI development:
- Independent third-party monitoring of AI models during their development phase
- Comprehensive regulations spanning across the entire technology industry
- Global regulatory frameworks to oversee cross-border deployment
Anthropic previously disclosed that it had identified and disrupted attempts to misuse its AI models for malicious activities, including potential support for biological weapons development. Meanwhile, the starkest warnings came from two safety team employees who resigned over safety concerns regarding machines surpassing human intelligence.
The debate over slowing down AI development highlights a profound tension within the tech sector between rapid commercial scaling and existential safety risks. As leading labs approach critical capability thresholds, industry calls for self-regulation and external oversight reflect growing anxiety over governance gaps.
Stock photo for illustration only, not from the actual event
In a separate interview with Fortune Magazine, OpenAI CEO Sam Altman noted that current safety standards are not yet equipped to push AI capabilities much further, conceding that advanced artificial intelligence operating beyond human control is entirely possible.
"I believe that if slowing down bought us even an extra year or two before models reach critical levels of capability, and we used that time to advance alignment, we could greatly reduce the risk that something goes seriously wrong"
Dario Amodei
Reactions across the industry varied widely. Clement Delangue, CEO of Hugging Face, launched the Open Alignment Initiative to support embedded evaluation concepts, while tech entrepreneur Elon Musk voiced his agreement on social media by stating that Dario is correct.
Source: BBC World
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment