Event Recap: Artificial Intelligence Implementation Case Studies
Why care about AI? DGI’s poll revealed that “60% of its audience is currently working on AI initiatives.” With case studies from the U.S. Patent and Trademark Office, Carnegie Mellon University and NIST, a number of experts addressed the relevance of AI and how to effectively implement initiatives.
Scott Beliveau with USPTO shared that his agency “looks at about a million trademark applications a year.” From his perspective, “AI helps get the next product or innovation to market.” Tom Scanlon with CMU provided a “unique perspective” because he is “seeing AI implementation across projects, multiple domains and parts of the government, and indicates that “there are common challenges that come up again and again. So, as you work on your own projects, you can identify the challenges and how to address them.” Lastly, Elham Tabassi with NIST walked through how they, “in collaboration with the community, developed a recently released AI Risk Management Framework” that is available in their Risk Management Playbook.
Specific topics included:
- AI implementation journey and scale from US Patent and Trademark Office
- metrics for success
- difference between using CONOPS and system requirements
- how to determine whether to stay on premise or go on the cloud with AI
- six common AI challenges across multiple projects at Carnegie Mellon University
- NIST’s 2021-2023 Roadmap of building the AI Risk Management Framework (AI RMF 1.0)