SXSW 2019
Trust Challenge: Why Explainable AI is not Enough
Description:
The success of artificial intelligence in the enterprise has prompted a great deal of discussion and debate around the notion of explainable AI, with transparency becoming the latest trait of trust. While transparency is a fundamental requirement, it actually has little bearing on the problem of trust. What the enterprise needs is justification, something entirely different than transparency. Dr. Jennifer Kloke will break down the difference between the two and discuss the importance of understanding why the results matter vs. how the results are achieved. Pulling from her work with Reliance, HSBC, Intermountain Health, Lockheed & others, she will explain why justification is required to move AI into production and what emerging techniques are moving us beyond explainable AI/transparency.
Other Resources / Information
Takeaways
- Why justification is the core element for success in artificial intelligence and is required for enterprise grade problems
- The difference between transparency and justification, and how the two together can provide more accurate and optimal results for the enterprise
- How to use methods like topological data analysis to justify an AI model’s behavior
Speakers
- Jennifer Kloke, VP of Product Innovation, Ayasdi
Organizer
Abrisham Khosravyani, Speaker Coordinator, credPR
SXSW reserves the right to restrict access to or availability of comments related to PanelPicker proposals that it considers objectionable.
Add Comments