The Biden-Harris administration has taken another step toward ensuring responsible development of artificial intelligence (AI) technology by securing voluntary commitments from major AI companies.
As part of the new effort, Amazon, Anthropic, Google, Inflection, Meta, Microsoft and OpenAI have committed to prioritizing safety, security and reliability in their AI systems.
Both companies are committed to several basic principles to protect the rights and safety of Americans. They conduct internal and external security testing of AI systems prior to release with the help of independent experts to prevent biosecurity, cybersecurity and broader societal risks.
In addition, these companies will share information on managing AI risks with industry, government, civil society, and academia to foster collaboration and best practices. They invest in cybersecurity and insider threat safeguards to protect the weight of their proprietary undisclosed model, releasing them only when it is intentional and safe.
To gain public trust, the two companies will also develop technical mechanisms, such as watermarking to indicate AI-generated content, to reduce the risk of fraud. It also reports on AI system capabilities, limitations, good and bad use, and covers security and societal risks, including fairness and bias.
Read more about AI and security: Google announces framework to protect generative AI
“Thoughtful regulation, ethical development, collaboration and high security practices allow us to harness the potential of AI for defensive cybersecurity purposes while minimizing its exploitation for malicious activity,” said Dave Randleman, CISO for Application Security and Ethical Hacking at Callfire.
“Effective regulation will enable cybersecurity professionals to use AI as a powerful tool to defend against cyberthreats and ensure a safer digital environment.”
The Biden-Harris administration said it is committed to ensuring U.S. leadership in responsible AI innovation. Alongside the enactment of the executive order, the administration will push for bipartisan legislation to safely manage AI development.
The administration is also working with international allies and partners to establish a robust global framework for AI governance.
“From a cybersecurity perspective, promoting guidelines may be effective for commercial gain, but it does nothing to deter threat actors from using technology to their advantage,” commented Mike Parkin, senior technical engineer at Vulcan Cyber.
“Voluntary schemes work within the ‘community’, but only in legally operated organizations.”
The new effort comes just days after the Biden-Harris administration announced the launch of the “U.S. Cybertrustmark” program on July 18.
