From left to right: Workday Senior Manager, Public Policy Evangelos Razis; Adobe Director of Ethical Innovation Grace Yee; BSA Managing Director Kate Goodloe; Box Vice President, Legal, Chief Privacy Officer & Global Head of Public Policy Leah Perry; and Microsoft Director, US Government Affairs Danyelle Solomon.
The development of responsible artificial intelligence (AI) is key to ensuring that there is public trust and confidence in AI technologies. This week, BSA | The Software Alliance welcomed industry stakeholders and over 60 Senate Staffers for a panel discussion on how companies are managing AI risks, including why tools like the NIST risk management framework (RMF) are important elements in AI legislation.
BSA’s Kate Goodloe moderated the panel of BSA member company representatives. She was joined by Box Vice President, Legal, Chief Privacy Officer & Global Head of Public Policy Leah Perry; Adobe Director of Ethical Innovation Grace Yee; Microsoft Director, US Government Affairs Danyelle Solomon; and Workday Senior Manager, Public Policy Evangelos Razis.
Panelists discussed how their companies are at the forefront of the development of responsible AI and detailed how they are accounting for the technology’s unique opportunities as well as its risks. The NIST RMF was highlighted as an important tool to help companies mitigate risks. Companies also spoke about the importance of using impact assessments in identifying and addressing risks associated with high-risk AI systems. All of the companies represented on the panel explained how their companies use tools like comprehensive AI ethics programs, AI principles, and other risk management tools to protect customer data and securely integrate AI capabilities into their products and services.
“The shared responsibilities between developers and deployers are dependent on different levels of control and visibility,” Razis said.
Yee echoed a similar sentiment, identifying Adobe’s AI governance program as a tool that uses internal ethics assessments to apply to vendors as well. The governance program is in addition to impact assessments, which are used in other sectors.
“These assessments are necessary to ensure that these technologies are not being used to make consequential decisions,” Perry said. She also emphasized how companies have established processes around data privacy impact assessments, which can be leveraged to conduct AI impact assessments.
Solomon spoke about how human-centric risk management programs are important to make sure that there is always a person making the final decision. Yee emphasized the importance of cultivating diverse risk management teams and focusing on how a company uses AI, to create more effective methods of identifying and mitigating risks.
The panelists highlighted examples of state and federal legislation that promote the use of risk management frameworks. They encouraged policymakers to focus legislation on circumstances where companies design or deploy AI systems used in consequential decision-making. They highlighted examples of state and federal legislation adopting that approach, including in the procurement context.
BSA continues to encourage the responsible development and deployment of AI by working with policymakers on efforts to promote AI risk management.
