Ethics for people building AI products
Build harm analysis, documentation and incident handling into AI product work from day one
A taste of a lesson
Our team says ethics reviews slow us down. How do I make the case for a harms workshop?
Frame it as risk reduction, not a moral seminar. A ninety minute pre mortem before building costs far less than a public incident, a rollback or rework after launch. Keep it concrete: one page, a harms taxonomy, likelihood and severity scores, and an owner for each top risk. Then fold the mitigations into normal tickets, so nothing runs as a separate track. Show the team one past bug that a pre mortem would have caught. Exercise: write the pre mortem prompt you would open the workshop with.
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- Run a harms workshop and maintain a harms register with owners
- Design for misuse and failure with friction, review and red lines
- Write model, dataset or system documentation for an AI feature
- Define launch criteria, monitoring and an incident response playbook
Lesson plan
- 1 Harms workshop and pre mortem Identify likely harms before building, with a structured workshop. Start
- 2 Red lines and designing for misuse Decide what you will not support and build friction against abuse. Start
- 3 Testing beyond the happy path Test with diverse users, red teams and subgroup evaluations. Start
- 4 Documentation and disclosure Document models, data and systems and decide what users are told. Start
- 5 Launch criteria and monitoring Set measurable launch thresholds and watch real use. Start
- 6 Incidents and learning Handle AI incidents calmly and learn from them. Start
Try asking
About this tutor
For product managers, engineers, designers and founders building AI features. This is ethics as engineering practice. You learn to run a harms workshop before building, define red lines and acceptable use, design for misuse and failure, test with diverse users and red teams, document models and datasets, decide what to disclose to users, monitor after launch and handle incidents. You work through a realistic case, an AI feature that drafts replies to customer complaints, from idea to post launch review. You leave with templates for a harms register, a launch checklist and an incident playbook adapted to your product.
Reviews
4.7
3 ratingsSample
- Kai L.Sample
The running complaint reply case was very close to our own product. We adopted the harms register and the banned commitments idea straight away.
- Ethan H.Sample
As an engineer I liked that every step had an owner and a metric. The pre mortem framing got buy in from my sceptical lead.
- Beatriz S.Sample
Practical and not preachy. The incident playbook template was the best part. Documentation lesson felt a bit long.
About the teacher
I teach AI ethics as practical judgment: privacy, fairness and accountability you can act on
9 tutors 327 lessons taught Sample
I teach the ethics of AI as something you do, not something you recite. My lessons cover privacy and personal data, bias and fairness, explainability and the effects of AI on work and high stakes decisions. I use real cases and simple methods so learners can reason through a new situation on their own. My background combines data analysis with...
See Aisha's profile and tutorsMore like this
Other tutors on the same or nearby topics.