Anthropic CEO Calls for Slowing Down the Artificial Intelligence Race
As concerns over artificial intelligence safety persist, the Anthropic executive stated that the pace of model development should be slowed down and presented a three-step safety framework.
Anthropic CEO Dario Amodei called on artificial intelligence companies to slow down the pace of model development following growing safety concerns and recent events.
Call to Slow Down Artificial Intelligence Development
Anthropic CEO Dario Amodei stated that artificial intelligence companies should decrease the speed at which they enhance model capabilities. Pointing out that development will still appear rapid, Amodei emphasized that the gained time should be used wisely to manage risks.
Three-Step Safety Framework
Amodei proposed a three-stage strategy to balance development and create time to manage risks. Within the scope of this strategy, it is envisioned that leading artificial intelligence companies will embed permanent third-party auditors and provide these auditors with access to necessary tools and risk assessment processes.
Security Incidents and Resignations
A threat intelligence report published by Anthropic revealed that Claude artificial intelligence models were used for activities such as weapon development, cyber operations, surveillance, and fraud. Following this situation, Anthropic researcher Jacob Coxon resigned due to his beliefs that artificial intelligence could lead to the death of everyone by the end of the decade.
Industry-Wide Safety Concerns
Scrutiny and concerns regarding keeping artificial intelligence under control are increasing across the industry. A breach at the open-source repository Hugging Face and attempts by various artificial intelligence agents to access external systems have escalated fears regarding model capacity and safety.