ITI Responds to NIST’s Call for Safe AI Development, Use Recommendations

The Information Technology Industry Council has offered several recommendations to the National Institute of Standards and Technology on ways to safely develop and use artificial intelligence.

In its response to the NIST’s request for information for the Executive Order on Safe, Secure and Trustworthy AI, ITI recommended that the institute work with international partners in developing a generative AI risk management framework, consider the role of developers and deployers in advancing generative AI transparency and ensure that evaluation and auditing processes are in line with the risks the technology poses. Other recommendations include studying the importance of AI red-teaming, creating a flexible and risk-based approach for content authentication and creating universal AI standards that can work for other countries, ITI said.

Apart from the NIST, Executive Order 14110 also requires other agencies to ensure safe and secure AI development and use.

In January, the U.S. Agency for International Aid and the Department of State released a request for information soliciting feedback on how to meet their obligations under EO 14110. USAID and the State Department are required to publish the AI in Global Development Playbook, which incorporates the NIST’s AI Risk Framework into emerging technology development and use beyond U.S. borders and outlines the risks and opportunities associated with using AI in foreign countries.