Trump signs executive order for voluntary AI model testing by government before public release

US President Donald Trump signed an executive order on Tuesday that allows leading artificial intelligence developers to voluntarily submit their most powerful models to the government for cybersecurity testing before releasing them to the public. The order represents a significant step in cooperation between the technology industry and the US government on AI safety issues.

The new regulation enables companies such as OpenAI, Google, and Anthropic to provide the government with access to their most advanced AI systems for up to 30 days before their planned release. During this testing period, government agencies can examine the models for potential security vulnerabilities and cybersecurity risks. Participation in this program remains entirely voluntary for companies.

Reports indicate that the initiative was triggered by concerns over Anthropic's Mythos model. The company had refused to release this particular model due to its exceptional capabilities in identifying and exploiting vulnerabilities in computer systems. This situation highlighted the growing challenges in managing increasingly powerful AI systems and their potential security risks.

The executive order directs various government departments, officials, and agencies to secure agreements with AI developers to test their models. Cybersecurity aspects are expected to be the primary focus of these tests. The testing aims to help identify potential threats before AI systems become available to the general public.

The voluntary nature of the program distinguishes it significantly from mandatory regulatory approaches common in other areas of technology regulation. Instead, the Trump administration is focusing on cooperation and incentives for the industry. Companies can decide for themselves whether they want to make their models available for government testing.

The order comes at a time when the debate over AI safety and regulation is intensifying internationally. Experts are increasingly warning about the potential risks of advanced AI systems, particularly in areas such as cybersecurity, disinformation, and autonomous weapons systems. At the same time, technology companies emphasize the importance of innovation and warn against excessive regulation.

For the affected companies, the program offers both opportunities and challenges. On one hand, they can strengthen trust in their technologies through cooperation with the government and identify potential security problems early. On the other hand, they must share sensitive technical details of their most advanced systems with government agencies.

The implementation of the order will reveal how many companies are willing to participate in the voluntary testing program. Industry observers expect that larger companies with established government relationships will be more inclined to make their models available for testing, while smaller startups may be more hesitant.

The initiative was reportedly prompted by specific concerns about AI models that could pose cybersecurity risks. As AI systems become more sophisticated, their potential to be misused for malicious purposes has become a growing concern for national security officials. The voluntary testing framework aims to address these concerns while maintaining industry innovation.

Government agencies involved in the testing will likely include cybersecurity experts and national security personnel who can evaluate the potential risks and benefits of new AI technologies. The 30-day testing window provides sufficient time for thorough evaluation while not significantly delaying product launches.

Long-term, this initiative could serve as a model for other countries developing similar approaches to AI regulation. The balance between innovation and security remains a central challenge for governments worldwide as AI technology continues to advance and create new possibilities as well as risks. The success of this voluntary approach may influence future regulatory strategies both in the United States and internationally.