Skip to main content

AI & Data Ethics

The most common path in our community. Companies building AI need people who can reason carefully about fairness, harm and responsibility, and philosophers are trained for that.

The work ranges from advising engineers on a specific system to writing the policies a whole company follows. What matters day to day is conceptual precision (what does “fair” mean in this system?), evaluating arguments, and translating principles into decisions other people can act on.

The lay of the land
#

Four kinds of employer hire here. Big tech companies have responsible-AI and policy teams sitting alongside the engineers. Consultancies and auditors check other companies’ AI systems. Policy institutes and non-profits work on the rules. And startups sometimes hire one person to cover all of it. Roles are advertised as AI ethics researcher, responsible AI analyst, AI policy analyst, AI governance manager or trust & safety analyst. Fresh graduates usually get in through research-assistant roles, fellowships, or volunteer communities like ForHumanity, which count as experience. A sensible start: take a free course, follow the field’s news, and publish something about the ethics of one concrete system.

Philosophers who’ve done it
#

Geoff Keeling
Bioethicist at Google
PhD in Philosophy, University of Bristol
  • Alessandra Fassio — our interview about getting into AI and data ethics (Part 1)
  • Katie Evans — our interview about working in AI and data ethics (Part 2)
  • Ravit Dotan, PhD — UC Berkeley philosophy PhD who advises startups and investors on AI ethics; spoke at our July 2022 workshop

From our events
#

  • From Philosophy to AI Ethics with Ravit Dotan (July 2022)
  • Bioethics and Ethics Consulting for Tech Companies with Geoff Keeling (November 2021)
  • Philosophy in the AI Industry (May 2021) and AI Ethics (April 2021)

Start here
#

Get the next story in your inbox

1,800+ philosophers get the Let's Phi newsletter: career stories, opportunities and resources for life beyond academia. It only goes out when there is something worth reading.

Prefer to browse first? Read the newsletter on Substack.