The United States, the United Kingdom, and 16 other nations have forged a landmark agreement aimed at securing artificial intelligence (AI) from the threats of malicious actors. Released on November 27, the 20-page document provides guiding principles that encourage the development and utilization of AI in a manner that prioritizes public safety and customer security.
Secure AI by Design
Focusing on the concept of “secure by design,” the agreement emphasizes the importance of integrating security measures throughout the AI design process. Jen Easterly, director of the U.S. Cybersecurity and Infrastructure Security Agency, told Reuters about the significance of aligning on the notion that AI systems should prioritize safety from the outset.
The UK government announced the guidelines alongsidethe first global guidelines to ensure the secure development of AI technology. The guidelines encourage companies to rigorously monitor AI for potential abuse, protect data integrity, and thoroughly assess software suppliers' security practices. While the agreement is non-binding and primarily offers broad recommendations, it marks a substantial step towards international cooperation on AI security.
Beyond Security: Regulatory Landscape and Challenges
The collaborative framework addresses how to defend AI technologies against hacking attacks and suggests that AI models should only be released following satisfactory security evaluations. However, the document does not engage with deeper issues such as the ethical deployment of AI or the processes involved in data collection for AI models. These concerns, ranging from democratic disruption to fraud and potential job displacement, remain salient as AI's influence grows within society.
European nations are currently leading the charge on AI regulation. France, Germany, and Italy have reached a consensus supporting mandatory self-regulation through codes of conduct, particularly for foundational AI models that generate diverse outputs. In contrast, the United States is still waiting for progress on AI legislation, despite initiatives from the White House aiming to mitigate AI risks and promote national security.
This international effort mirrors several previous endeavors to shape the trajectory of AI development globally, though many past initiatives lack enforcement mechanisms. As the world increasingly grapples with the pervasive impact of AI, this agreement sets a benchmark for future discussions and potential regulations in the AI domain.
