Responsible AI: Rewriting or Repeating History?
summary
History shows concrete failures and fixes: photographic tech takes 110 years until color photos serve darker skin, car safety goes from red flags to airbags over decades while first female crash test dummy appears 50–60 years after airbags, self‑driving systems cause a pedestrian death, pulse oximeters misread darker skin, LLM image models sexualize an Asian user, clickworkers face abusive conditions, and one estimate ties a ChatGPT query to ~1.5 L of water; regulators get minutes, not decades, to react and industry urgently needs scalable responsible‑AI tooling and audits.
AI-written summary · about
chapters (17)
- 0:00Responsible AI keynote intro
- 0:47History’s tech failures lesson
- 3:17Photography’s skin tone bias
- 4:44Regulators’ shrinking timelines
- 7:58Defining responsible AI
- 10:07Tech safety track record
- 14:02AI societal polarization
- 15:48Digital divide challenges
- 17:35Generative AI bias ethics
- 21:47Red-teaming AI harms
- 22:52AI jobs vs productivity
- 23:55AI colonialism exploitation
- 26:50AI environmental cultural costs
- 28:58Responsible AI collective habit
- 31:49Scalable responsible AI ecosystems
- 33:54Inclusive AI adoption steps
- 34:16Q&A on dark AI future
related talks
34mResponsible AI: The Key to Ethical, Safe and Inclusive Software Development
35mResponsible AI: Ethics in Software Development
23mThe AI Dilemma — Steering towards a More Humane Future
40mSocially responsible AI: Practical tools for robust design
29mResponsible AI Challenges and Recommendations
44mPanel: Building responsible, human-centric AI models
38mBuilding Responsible AI in the Age of GPT
24m2025: year of Ai dilemma - ethics, regulations and innovations