Interview with Sabrina Shih, AI Policy Integration Lead at Responsible AI Institute

Could you tell us a bit about the Responsible AI Institute?

"Founded in 2016, Responsible AI Institute is a global and member-driven non-profit dedicated to enabling successful responsible AI efforts in organizations. We accelerate and simplify responsible AI adoption by providing our members with AI conformity assessments and guidance closely aligned with global standards and emerging regulations."

Could you summarise your journey with Generative AI thus far?

"We have worked closely with large companies deploying Generative AI across their operations and have observed the challenges and best practices that underlie responsible development, operation, and use. The technology moves quickly, and we are closely following new model releases and new regulatory developments."

Could you summarise the US policy and regulation surrounding GenAI for us?

"The implications of US policy and regulation on GenAI are still materializing. Federal agencies have been addressing certain aspects of GenAI, such as FTC’s comment late 2023 noting its commitment to protecting copyright and consumers against violations by GenAI developers and deployers. Earlier this year, NIST published its GenAI Profile, a GenAI specific extension of its NIST AI Risk Management Framework. The US AI Safety Institute (and its UK counterpart) is expected to develop and introduce new evaluations for AI models and standards for content authentication, among other potential elements for regulatory oversight, although these may be more relevant to large frontier model developers."

What have been your/orgs you work with’s biggest challenges?

"With respect to GenAI, our member organizations often already have a good idea of how they hope to utilize GenAI capabilities. Their commonly noted challenges instead focus on understanding whether risks from their systems have been sufficiently managed, whether their systems are made fit-for-purpose for their deployment contexts, and whether most known responsible AI objectives with respect to systems (e.g., fairness, privacy, reliability) have been meaningfully addressed." 

How are you overcoming these?

"As we conduct assessments with our member organizations, we serve as a second pair of eyes to review the organization’s governance, policies, and process and evaluate them against authoritative guidance on these topics. Working closing with member organizations, we learn together what a responsible GenAI approach looks like at both the enterprise and system levels. "

Are there any tangible results/successes/metrics you can share?

"Through our Organizational Maturity and System Assessments, our member organizations have built controls for Responsible AI, strengthened relevant principles and partnerships, and operationalized cross-department collaboration to accelerate their GenAI strategy. Case studies on this work can be found on our site."

Are there any potentially surprising/unexpected changes you made/things you have learned/implemented that have made a big difference to your GenAI efforts that you might recommend to others?

"Have a conversation with your peers on what they mean when they say Generative AI! The difference in conception may be surprising, even within intimate teams, and reflects the varied conversations across enterprise and technical landscapes. Aligning on a standard definition of GenAI (e.g. rooted in specific capabilities, outputs, algorithms, etc.) is a foundational step to any enterprise GenAI strategy."

Can you give us a teaser for your talk at the upcoming GenAI Week?

"I am excited to be hosting a deep dive on GenAI use case development and maintenance from a Responsible AI lens. How can we make informed decisions when expanding or adapting systems to new contexts or purposes while still prioritizing performance, safety, and transparency?"

Why are you looking forward to the conference this October?

"I look forward to connecting with the amazing assembly of industry experts at the conference and learning how AI governance, risk management, and safety considerations compare across their profiles."

Want to join Sabrina at the GenAI Week? Download the agenda here and see what sessions are lined up for this event

Return to Home