Chính quyền Trump yêu cầu OpenAI trì hoãn phát hành GPT 5.6, chỉ chia sẻ với đối tác thân cận, do lo ngại an ninh mạng và lạm dụng công nghệ.
OpenAI’s New Approach with GPT 5.6 Release
OpenAI’s strategy for releasing its latest language model, GPT 5.6, marks a significant departure from its previous practices.
Restricted Access to GPT 5.6
Unlike its earlier models, OpenAI intends to limit the availability of GPT 5.6 to a select group of partners, rather than making it publicly accessible. This decision was reportedly influenced by the Trump administration’s guidance, indicating a shift in the company’s approach to model distribution.
Government Involvement and Approval Process
According to The Information, OpenAI’s CEO, Sam Altman, informed employees at a meeting that the government would play a direct role in the release process. The administration will approve access to GPT 5.6 on a customer-by-customer basis during an initial preview period. This level of government involvement in AI model distribution is noteworthy and suggests a growing interest in regulating AI technologies.
Potential Broader Release
OpenAI has indicated that if the limited release proves successful, they aim to follow up with a more general release a couple of weeks later. This two-step approach allows them to gauge the model’s impact and address any potential issues before making it widely available.
Comparisons with Anthropic’s Strategy
OpenAI’s new release strategy draws parallels with Anthropic’s handling of its powerful AI model, Claude Mythos.
Anthropic’s Limited Release of Claude Mythos
Anthropic made headlines earlier this year by announcing that its advanced cyber model, Claude Mythos, would be released exclusively to a small group of partners through Project Glasswing. This decision sparked debates about the motivations behind such a restricted release.
Debates Surrounding Anthropic’s Approach
Observers have questioned whether Anthropic’s limited release strategy is a genuine effort to prevent misuse of a powerful model or merely a marketing tactic. The argument revolves around the model’s potential to cause harm in the wrong hands, given its capabilities. While some view this as a responsible approach, others see it as a way to generate buzz and interest.
The Middle Ground
The truth might lie somewhere in the middle, as Anthropic’s concerns about misuse are not entirely unfounded. With the rise of generative AI, cybercriminals have access to more sophisticated tools. LLMs have demonstrated their proficiency in creating malware, and some can even autonomously execute ransomware attacks. Models like Mythos, which can identify and exploit software vulnerabilities rapidly, raise legitimate security concerns.
Government’s Growing Interest in AI Regulation
The Trump administration’s involvement in OpenAI’s release strategy is part of a broader trend of increasing government oversight in AI development.
Shift in Government Stance
Initially, the Trump administration advocated for a hands-off approach to AI regulation. However, recent actions indicate a change in direction. The administration has begun pushing for federal oversight of new AI models, as evidenced by the executive order signed by President Trump earlier this month.
Executive Order and Federal Oversight
The executive order directs specific AI companies to voluntarily submit their new models to the government for testing and evaluation before public release. This move is a significant step towards establishing federal control over AI technologies and ensuring they are thoroughly vetted before widespread adoption.
Addressing Security Concerns
The government’s interest in regulating AI models is likely driven by security concerns. As AI technology advances, the potential for misuse and abuse increases. By implementing oversight, the administration aims to mitigate risks associated with powerful AI models, especially those capable of identifying and exploiting software vulnerabilities.
Balancing Innovation and Regulation
The challenge lies in striking a balance between fostering AI innovation and ensuring public safety. While government oversight can help address security risks, it should be carefully designed to avoid stifling technological progress and maintain the competitiveness of AI companies in the global market.
