DeepSeek Shakes Global Balance, Reignites Regulatory Debate
As China advances in generative AI, there's a growing global consensus on the need to regulate these powerful models. However, cooperation between the United States and China presents a significant challenge.
Until recently obscure, the Chinese AI startup DeepSeek has stunned the global community by releasing one of the most advanced **open-source generative AI models**. As AI capabilities in China expand, so do the risks associated with powerful AI models.
## A Race Toward Open AI with Its Challenges and Risks
A major challenge now confronts AI companies. Both the American and Chinese governments are struggling to regulate AI due to its large-scale risks. Despite these concerns, a surprising consensus has emerged among major AI developers in both countries on the need for safeguards, including DeepSeek.
These rapid advancements in model capabilities have been accompanied by growing concerns **”about the risks”**, often highlighted by the companies themselves.
## The Rise of an International Consensus on AI Regulation
AI summits have nurtured an industry-wide consensus to mitigate risks. For instance, at last year’s Seoul AI Summit, several companies made commitments now known as the **”Seoul Commitments”**. These efforts aim to establish best practices to prevent catastrophic risks that could arise with increasingly powerful AI.
Despite calls for international consensus, coordination between the US and China on frontier AI has seen little progress. However, the Chinese industry is joining the global push for AI safety commitments, which is particularly significant.
## A Convergence of Commitments on AI Safety
In this light, the ongoing AI Action Summit in Paris could mark a turning point for the industry in terms of consensus on frontier AI risks. If more Chinese companies join the Seoul Commitments and publish safety frameworks alongside their Western counterparts, it could inject new vitality into industry-led safety commitments as a pathway to global risk mitigation.
While the US and China find it challenging to forge a governmental agreement, companies on both sides of the Pacific might build a temporary scaffold for governing frontier AI until governments establish a permanent structure.