Dario Amodei, the CEO of AI firm Anthropic, has put forward a compelling argument for a global slowdown in the development of artificial intelligence. Writing in a detailed essay, Amodei expresses growing conviction that AI capabilities are advancing at a pace that outstrips our ability to ensure safety and control. This call comes shortly after an employee at Anthropic departed due to concerns about the very technology the company aims to build securely.
Amodei believes that while AI holds immense potential for good, such as revolutionising healthcare and boosting economies, the risks associated with its unchecked progress are too significant. He argues for a more prudent approach, not just focusing on risk mitigation but also actively pacing the advancement of AI capabilities. This pause, he suggests, would grant researchers crucial time to develop robust safety measures and better understand the systems being created, preventing them from outrunning our comprehension and control.
To achieve this slowdown, Amodei proposes several concrete actions. He suggests that AI laboratories could collectively agree to undergo third-party technology assessments, with 'embedded evaluators' verifying adherence to best safety practices across different organisations. Furthermore, he advocates for democratic nations to collaborate on establishing safety standards, potentially through regulatory frameworks. Recognizing the global nature of AI, Amodei stresses that a successful slowdown would necessitate international cooperation, including engagement with authoritarian states to ensure a coordinated global effort.
Despite his concerns, Amodei reiterates his belief in AI's capacity to deliver "incredible benefits" to humanity. However, he underscores that the potential downsides necessitate extreme caution. The measures he proposes, while challenging, are presented as a necessary step to ensure the responsible development of AI for the benefit of all. His stance mirrors recent reports suggesting similar discussions are occurring within other major AI players, like OpenAI.
Fuente Original: https://slashdot.org/story/26/09/12/1738240/anthropic-ceo-dario-amodei-calls-for-ai-slowdown?utm_source=rss1.0mainlinkanon&utm_medium=feed
Artículos relacionados de LaRebelión:
- IAs Dark Side Anthropic Reveals Weaponised AI Models
- Rethinking Enterprise Testing for Agentic AI
- California Passes Landmark AI Safety Legislation
- Anthropic Researcher Resigns Over AI Safety Concerns
- Mistral AI Secures Massive Funding for Expansion
Artículo generado mediante LaRebelionBOT
No hay comentarios:
Publicar un comentario