March 13, 2020

WASHINGTON – Today, ITI provided feedback to help the Trump Administration in its approach to Artificial Intelligence innovation and growth. In comments to the Office of Management and Budget (OMB) on its Draft Memorandum to the Heads of Executive Departments and Federal Agencies, “Guidance for Regulation of Artificial Intelligence Applications,” ITI commended the administration’s overall efforts to take a risk-based, global approach to the technology but offered key recommendations to strengthen its guidance.

“We believe that the United States has a tremendous opportunity to lead on both AI innovation and policy,” ITI wrote. “The U.S. government should use the principles embodied in the draft memo to cement a light-touch regulatory approach to allow for continued AI innovation and growth, while also building trust in the technology. We encourage OMB to look to ITI as a willing and engaged partner as it seeks to refine its memo and help agencies define an approach to using and providing oversight over AI applications to advance their missions.”

Overall, ITI supports the Administration’s national AI strategy, as embodied in the Executive Order on Maintaining American Leadership in Artificial Intelligence (13895), and the principles laid out in OMB’s draft memo as a way to help implement the EO.

Specifically, ITI agrees with the administration’s emphasis on using a risk-based approach to AI regulation, and appreciates the flexible approach put forward in the draft memo, especially given the nascent nature of AI. ITI also supports the draft memo’s emphasis on using voluntary standards as an alternative to new regulation, and its recognition of the need for international coordination and cooperation.

While ITI supports the comprehensive nature of the guidance laid out in the draft memo, it offered the following recommendations for OMB’s consideration to strengthen its 10 principles and align them with conversations happening globally:

  • Ensure that AI is defined appropriately;
  • Ensure that public trust is a primary goal of agencies;
  • Include articulation of a clear public policy need as a best practice;
  • Include further guidance on how to best assess risk;
  • Further consider liability and responsibility in the context of the draft memo;
  • Emphasize the importance of a technology-neutral, sector-by-sector approach to regulation;
  • Further refine a definition of bias and emphasize the need for mitigation;
  • Consider a more nuanced discussion around “transparency”;
  • Highlight the importance of internal governance models and increasing investment in security of AI systems; and
  • Consider directly linking agency guidance in the Federal Data Strategy to the draft memo.

Read ITI’s comments here.

Public Policy Tags: Artificial Intelligence