Ben Byford
AI ethics | freelance web and games designer
A tiny, weenie, short, but perfectly formed introduction to AI Ethics
What is right to do?
What systems or frameworks should we make to help?
What behaviour is and isn't permissible or desirable?
What kinds of opportunities are we making available or not?
What does a good life look like?
What ought we create?
What direction are we taking the world in?
wikipedia
Ethics or moral philosophy is a branch of philosophy that involves systematizing, defending, and recommending concepts of right and wrong conduct.
Laws
Money / Grants / Tax
Standards / frameworks / code of conduct
Culture / cultural artefacts / traditions
Bureaucracy / Systems / Computers
Jaron Lanier
knowledge erosion / de-skilling, inappropriate AI uses, total surveillance, AI value alignment, diverse workforce, human rights and measurements of human flourishing, destablising democracy / political exploitation, lack of accountability, job displacement, lack of inclusion, fairness / bias, impersonation / deepfakes, environmental impacts, unintended consequences, lack of transparency, black box / explainability, data protection, consent, intellectual property rights, data manipulation, whos ethical values, information obfuscation, duplication of personal data, security, safety, perversion, robustness, personhood, singularity, world without work, human value
Anderson, Anderson 2012
www.benbyford.com
www.benbyford.com
Product definition / Intention
Horizon scanning / consequence scanning
Speculative design / design fiction
Value sensitive design
Design Critique
Designated desenter
Archetypes
The Veil of Ignorance
Confusion matrix
Ethical Risk assessment
| Risks | Opportunitities | To whom? |
|---|---|---|
|
|
We're starting a business!!!
product / service
- Product name
- What it does
e.g. Artificially intelligent robots for dogs.
product / service
product / service
2yrs
indirect
direct
badly
well
consequence scanning:
it went well / it went Badly
product / service
2yrs
indirect
direct
20yrs
badly
well
consequence scanning:
not much, amazing, exploitative?
product / service
2yrs
indirect
direct
20yrs
50yrs
consequences
badly
well
consequence scanning:
Is the world a great place or a horrid dystopia?
consequence scanning:
Think about the people, animals or things it effected, the kinds of scenario's that might happen, and the actions it allows or disallows e.g. human rights.
consequence scanning:
What is the risk or opportunity?
To whom does it affect?
consequence scanning:
Generated idea
Product name
Most interesting futures
Ethical considerations
Alan Kay
www.benbyford.com
By Ben Byford
talk and workshops for Blackheath school 2025