IBM submitted the following comments to The White House regarding their recent memorandum on “Guidance for Regulation of Artificial Intelligence Applications” for the heads of executive departments and agencies. (Download the PDF letter here)

 

March 2, 2020

 

Mr. Russell T. Vought
Office of Management and Budget
725 17th St., NW
Washington, DC 20503

 

RE: Memorandum for the Heads of Executive Departments and Agencies, Guidance for Regulation of Artificial Intelligence Applications

 

Dear Mr. Vought:

 

IBM appreciates the opportunity to comment on the recent memorandum for the heads of executive departments and agencies regarding the issue of “Guidance for Regulation of Artificial Intelligence Applications.” As the world’s leading supplier of trustworthy AI systems and solutions, IBM has firsthand experience with the many possible commercial applications and societal benefits associated with this emerging field of technology.

 

We commend the White House’s continued focus on this area of technological governance and are pleased to see the administration’s balanced approach to addressing both the benefits and concerns associated with AI. This memorandum’s focus on the need to develop “a regulatory approach that fosters innovation, growth, and engenders trust, while protecting core American values” is an important step forward in the global dialogue surrounding governance of AI, and will help to position the United States as a continued leader in this field.

 

Overall, IBM is supportive of the direction represented by this guidance and offers the following recommendations for consideration:

 

• Recognize the different roles in the AI developmental lifecycle by further distinguishing between different applications of AI and the types of firms creating and deploying them;
• Prioritize greater international engagement through increased coordination and cooperation with global partners; and
• Ensure greater transparency of how agency plans are submitted, assessed, and approved, by acknowledging that such plans will be open to public comment or consultation.

 

Once again, IBM appreciates the opportunity to comment and we look forward to future engagements on this and other issues.

 

Sincerely,
Christopher Padilla
Vice President
IBM Government and Regulatory Affairs

 

IBM Response to “Memorandum for the Heads of Executive Departments and Agencies, Guidance for Regulation of Artificial Intelligence Applications”

 

IBM appreciates the opportunity to comment on the Office of Management and Budget’s (OMB) recently released draft of the “Guidance for Regulation of Artificial Intelligence Applications” (hereafter, “the guidance”). As a company, we have long supported the need for industry to prioritize the responsible development and deployment of new technologies. This guidance marks an important step in the journey towards a business environment where trusted AI systems are the norm. We thank OMB for the important recognition this topic deserves and believe that “Public Trust in AI” is appropriately situated as the first of these ten principles.

 

We are broadly supportive of this guidance and its inclusion of a wide range of important governance principles, from prioritizing trust and transparency to the necessity of cost-benefit analyses in agency rulemakings and risk-proportionate assessments of AI applications. But we also believe there are a number of ways the guidance can be strengthened. These comments examine three such areas in which there is a need to:

 

• Recognize the different roles in the AI developmental lifecycle by further distinguishing between different applications of AI and the types of firms creating and deploying them;
• Prioritize greater international engagement through increased coordination and cooperation with global partners; and
• Ensure greater transparency of how agency plans are submitted, assessed, and approved, and acknowledging that such plans will be open to public comment or consultation.

 

Recognize the Different Roles in the AI Developmental Lifecycle

 

IBM has been a leading proponent of the concept of “precision regulation” – the idea that rules and regulations should target specific harms and outcomes, not broad categories of technology. In a recent essay, the IBM Policy Lab made this point in the context of AI, arguing that “[g]iven the ubiquity of AI – it touches all of us in our daily lives and work – there will be no one-size-fits-all rules that can properly accommodate the many unique characteristics of every industry making use of this technology and its impact on individuals.”

 

Building from this, we note that an implicit recognition of the fundamental difference in accountability between stages of AI development can help appropriately assign responsibility for providing transparency and ensuring fairness and security, based on who has better control over the protection of privacy, civil liberties, and harm-prevention activities in a given context. … Risk-based regulatory approaches like this – which also allow for more manageable and incremental changes to existing rules – are ideal means to protect consumers, build public trust in AI, and provide innovators with needed flexibility and adaptability.

 

The guidance recognizes these nuances, embracing the need for a risk-based approach to determine “which risks are acceptable and which risks present the possibility of unacceptable harm” in the context of specific AI applications. However, the guidance should go further by recognizing the unique and distinctive characteristics of various types of organizations involved in the AI supply chain. For example, some firms may contribute to the research and development of basic AI models or the creation of tooling and APIs, while others specialize in the training, management, and ultimate control or operation of AI models in a commercial context. Generally speaking, this bifurcation can be drawn according to whether an organization is involved in the early-stage lifecycle design and development of an AI model (a “provider”), or whether it is engaged in later-stage operational deployment of the technology (an “owner”). These different organizational roles should be taken into account when agencies consider new rulemakings concerning this technology. We recommend the guidance create a more unambiguous reference to the need for agencies to take account of the different characteristics between providers and owners of AI systems when considering new rules and regulations.

 

The appendix to the guidance, for example, discusses the appropriate mechanisms for agencies assessing and managing risks when considering the need to pursue a regulatory versus non-regulatory approach to AI. Unfortunately, there is no mention of making determinations of risk or mitigating the potential for harm based on the type of organization utilizing, deploying, or managing the technology. As such, OMB should clarify that agencies considering future AI rulemakings should take account of those firm-specific differences, in addition to a particular AI application’s risk profile.

 

We recommend OMB make its directions to agencies clear by creating an additional section under “Appendix A: Technical Guidance on Rulemaking” (or elsewhere in the guidance) that focuses on the need for any future rulemakings to take account of an organization’s role in the larger AI supply chain ecosystem.

 

Prioritize Greater International Engagement

 

While we support the recommendation that directs agencies to “keep in mind international uses of AI” with regards to the global competition effects of any proposed AI rulemaking, it is important for federal agencies to recognize that there are already a number of broadly agreed upon international AI frameworks and principles. In addition, there are a number of countries, such as Brazil and Canada, that are currently working through comment and consultation periods as a means to developing their own regulatory approaches to AI.

 

Building on existing international engagement efforts will be an important cornerstone in working towards transnational harmonization of AI standards and regulations. In order for the US to help drive that dialogue, it is imperative for the administration to continue prioritizing conversations in multilateral forums, such as the OECD, as well as more focused and constructive engagements with the EU. Indeed, the administration has repeatedly recognized the importance of such efforts, most recently in its American Artificial Intelligence Initiative: Year One Annual Report, which notes, in part:

 

In all areas of strategic emphasis, partnerships and collaboration with academia, industry, nonprofit organizations, civil society, other non-Federal entities, and international partners and allies are of growing importance. … The Nation also benefits from relationships between Federal agencies and international partners who work together to address key challenges. AI partnerships with allies and partners represent one of our sources of strategic competitive advantage.

 

IBM concurs with this sentiment and believes that the best means for the US to continue leveraging this key “strategic competitive advantage” is to continue to invest in these international relationship-building efforts, especially in our relationship with the EU.

 

Striking a balance between regulatory approaches in the US and the EU will be a key determining factor in the future success of AI. To that end, IBM welcomes the general direction taken by the European Commission in its recent white paper on AI. Its overarching tone is aligned with both a precision regulation approach to AI governance, as well as with the risk-based framework proposed in the OMB guidance. Given the general similarities in both the US and EU approaches, we recommend the administration take this opportunity to show its leadership on AI by continuing to work through the OECD, and other international forums, towards achieving an international consensus on AI.

 

In support of such a dialogue, the guidance should also direct federal agencies with an international focus to develop a strategy for engaging with partners in the EU, East Asia, Latin America, and elsewhere. As such, we support the guidance’s recognition that “agencies should engage in dialogues to promote consistent regulatory approaches to AI that promote American AI innovation while protecting privacy, civil rights, civil liberties, and American values.” However, we believe these dialogues must necessarily include an international dimension to them. This guidance could be significantly strengthened by expanding on the need for federal agencies – when and where appropriate – to prioritize international engagement with other countries in order to minimize the potentially disruptive effects of incongruent international regulatory treatments of AI.

 

In directing agencies to submit plans to showcase consistency with the guidance, we recommend the guidance expand the “International Engagement” section to direct agencies to consider existing international frameworks to which the United States has committed itself (e.g., OECD AI Principles) and prioritize the development of strategic plans for coordination and cooperation with international partners. This will help promote the need for transnational dialogues in order to build upon the solid bedrock of international collaboration that the administration and international partners have already helped engender.

 

Ensure Greater Transparency of Agency Plans

 

Finally, we are pleased to see the guidance takes seriously the need for public participation and greater public-private collaboration. Prioritizing the need for public participation throughout any proposed rulemaking process is an important means of promoting consistency and transparency in governance practices. The need for transparency in regulatory rulemaking is not only noted in this guidance, but also in the administration’s October 2019 Executive Order on Promoting the Rule of Law Through Improved Agency Guidance Documents.

 

To that end, IBM recommends including an explicit recognition that these same principles will be adhered to when agency plans for achieving consistency with the guidance are submitted to OMB, and that all agency plans, and any OMB assessment of such plans, will be open to comment and consultation.

 

Conclusion

 

IBM applauds the administration for its commitment to engaging on this issue, and for taking the lead on promoting a regulatory approach that can help effectively balance innovation and the public interest. This is an important step in America showcasing its thought leadership on an important emerging technology, and we look forward to the administration’s plans for advancing the AI governance conversation not only here in the United States, but with partners from around the world.

 

Media Contact:
Sarah Minkel
sarah.minkel@ibm.com

 

Share this post: