Mozilla Festival 2021 (March 8th – 19th, 2021)

Mozilla Festival 2021 (March 8th – 19th, 2021)

Ethically Using AI to Watch TV

This session will be a discussion about the challenges of using new (and sometimes messy) algorithms to generate novel analysis that might shape opinions and understandings of power. When is it OK to track a face? Is it acceptable to analyze faces in the name of monitoring coverage? What if the algorithms are flawed? At what point does a highlight reel become misinformation? How much error should we tolerate?

  1. We will begin with a very brief presentation (<5m) to set the stage / context.
  2. The majority of the session will be a group discussion of specific ethical design challenges -- e.g. "You would like to understand whether women are given appropriate air time on your local news station, but gender identification algorithms are more likely to misgender women of color" -- participants would be asked to think through / design ethical boundaries together.

What is the goal and/or outcome of your session?:

Our goal is for participants to walk away with more awareness of the ethical challenges surrounding the use of algorithms to generate novel information.

We're hoping that many efforts and discussions will continue after Mozfest. Share any ideas you already have for how to continue the work from your session.:

We are building a community and set of open source tools dedicated to remixing / analyzing video and television using code (https://tv.kitchen).

We hope that the discussion in this session can be the start of a guidebook / handbook about ethical uses of AI within the TV Kitchen (and any other project that is using code / AI to analyze media and current events).

How will you deal with varying numbers of participants in your session?:

I would like to speak with a wrangler about this -- particularly with regards to what is possible via the platform being used for the virtual conference.

The idea of group discussions of specific scenarios can really work at any level, but ideally we could handle larger numbers of participants by breaking them into smaller groups and doing "share backs"

Dan uses technology to build things for the public good. He is currently a Corporate Overlord at Bad Idea Factory working on tools to help journalists track trends on TV.