Anthropic and OpenAI look to Uncle Sam to make them too big to fail
American model devs are trying to convince Washington to cement their dominance
Two leading AI companies, Anthropic and OpenAI, are seeking government support to become "too big to fail." Despite billions in investment, the firms are not yet profitable and rely on a steady stream of capital. Meanwhile, Chinese open-weight AI models are advancing rapidly, sometimes surpassing the capabilities of the leading US models while using fewer resources.
Concerns about the potential dangers posed by unchecked AI have gained traction after researchers discovered a single prompt could access sensitive data. This led to Anthropic and OpenAI complying with export control directives, showing the extent to which the companies are willing to bend to government pressure. The fear of AI's potential for harm has played into the hands of these companies, allowing them to push for increased regulation.
However, some experts argue that the issue is not the models themselves, but rather the inadequate safeguards in place. They point out that airgapping sensitive computing workloads, a practice known for decades, could have prevented many breaches. Instead of focusing on slowing down AI advancements, it may be more effective to address the weak points in current safety measures.
Written by urgent.news from The Register's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
