SXSW 2019

On Bots & Bias: When What Machines Learn Is Wrong

Description:

How do we humans build bots that don’t suck—especially when training data sources like Twitter or Reddit can corrupt them into “Hitler-loving sex robots in 24 hours”? And WTF is up with gendering AI assistants female? Join industry leading ladies for some real talk on building and scaling intelligent bots. We’ll debate hot button issues like when to use AI and ML—and whether or not it really works. Should we design conversational software based on UX and consumer preferences—which many cite as why we’d rather boss around Alexa than Alex—or do we build for a better society even if it means censoring data? Hear diverse perspectives from experts on the Enterprise, Developer, Data Analytics, and Brand Experiences to learn critical bot building dos and don’ts in this rapidly evolving ecosystem.


Related Media

Additional Supporting Materials

Other Resources / Information


Takeaways

  1. Behind every bot is a human -- how do we design these systems to not reflect the worst of human nature and human behavior on the internet?
  2. What is the current state of the art in Natural Language Understanding and Artificial Intelligence, and how effective is Machine Learning?
  3. How to build intelligent bots as a Brand, Enterprise, or Developer.

Speakers


Organizer

Anamita Guha, Product Manager, IBM Watson


Meta Information:

  • Event: SXSW
  • Format: Panel
  • Track: Intelligent Future
  • Track 2
  • Level: Intermediate


Add Comments

comments powered by Disqus

SXSW reserves the right to restrict access to or availability of comments related to PanelPicker proposals that it considers objectionable.