What You Will Learn
1: 10 Commandments of Responsible AI Use
Presented by Katie Hasty
As artificial intelligence becomes more integrated into business practice, the ethical risks — and expectations — are evolving just as quickly. Presenter Katie Hasty, CEO of Prima Vista, will guide attendees through emerging issues related to the use and abuse of Large Language Models, and provide a practical framework for evaluating ethical AI use in a variety of scenarios.
Through hands-on exercises, participants will work through realistic scenarios in which AI tools are misused, identifying key ethical breakdowns and discussing how to avoid similar pitfalls in day to day use. The program will also provide practical strategies for guidance implementation, along with a downloadable checklist attendees can use as an ongoing resource to support ethical, informed use of AI tools.
2: What AI Can and Cannot Do Right Now
Presented by Russ Fischer and Katie Hasty
Sweep away hype and marketing to learn what Large Language Models can realistically and responsibly do in a professional context. Responsible AI use requires an understanding of how AI works. We will break down the fundamentals of Large Language Models and Generative AI to highlight the aspects that make AI useful in a professional capacity, and those which can lead to serious liabilities.
This is not a technical session, but it does root down into the foundations of Large Language models to explore how problems such as hallucinations are baked into the technology. Attendees will learn why models deliver output in specific ways and how that output is generated — and therefore, how to structure prompts and queries to get the most out of the technology.
Afternoon: Two 85-min sessions, 10 minute break in between
1: Prompt Lab 1
2: Prompt Lab 2
Presented by Russ Fischer and Katie Hasty
Put the rules and background concepts discussed in the morning sessions to work in a fully interactive, hands-on approach to learning AI. Using a set of custom tools, we present prompting lessons in a structured format that minimizes typing downtime while allowing learners to engage with complex prompt problems. We encourage discussion and the comparison of results from different models.