Stream: Wellington A
Time: 12:15 - 13:00
With hallucination rates theoretically as low as 1% but practically as high as 50%, most AI vendors tout the "Human in the Loop" to catch errors. But what is that Human in the Loop actually doing? Are they reviewing thousands of lines of generated code? Are they double checking every loan decision, every transaction flagged as potential fraud? In this session, we'll review real examples of how to ensure AI reliability from Human in the Loop to AI reviewers. These techniques will be invaluable to anyone considering leveraging AI.
There is currently no attachment for The (im)Practical Reality of the "Human in the Loop"
Through 3 decades, 5 companies and 12 job changes, 2 things have remained constant in Misty's career: mainframes and a passion for helping people. She's currently the worldwide Tech Lead for the IBM business group at Accenture as well as the Chair of the OMP Modernization Working Group continuing her personal mission of helping companies modernize their existing mainframe applications. She's also known as Darth Misty the Mainframe Sith and quite possibly has the world's largest collection of mainframe shirts.
Computer Science student at VIT Chennai and an active contributor to the IBM Z and open mainframe ecosystem. Served as an IBM Z Student Ambassador since 2024 and has been recognised as an IBM Champion for 2025 and 2026. An Open Mainframe Project mentee working with Misty Decker on research exploring the role of AI in mainframe modernisation, with a focus on how emerging technologies can support the evolution of enterprise mainframe applications. Also founded the IBM Z Student Society India and co-founded Women in Z to help students connect with and explore opportunities in the mainframe community. Interests lie at the intersection of IBM Z, AI, mainframe modernisation, and open source, with a focus on learning, community building, and helping the next generation of students engage with mainframe technology.