Last Week in AI The Path to Facial Recognition Reform and Regulation
Jun 19, 2020
This discussion highlights critical concerns regarding facial recognition technology sold to law enforcement. Experts emphasize the urgent need for regulations to address privacy rights and racial justice. They explore the bias inherent in these systems, particularly against minorities and women. There's a strong call for corporate accountability to support racial justice initiatives. Additionally, ethical implications are examined, linking these technologies to historical injustices like eugenics. The conversation is a vital step toward responsible AI development.
AI Snips
Chapters
Transcript
Episode notes
Facial Recognition Backlash
- 70+ civil rights and research organizations demanded Amazon stop providing facial recognition tech to governments.
- Gender Shades project highlighted AI bias in gender classification tasks, influencing IBM, Microsoft, and Amazon's decisions.
Bias and Dangers of Facial Recognition
- Facial recognition systems exhibit significant bias, performing worse on darker-skinned women than light-skinned men.
- Even accurate systems pose risks due to potential misuse, prompting calls for bans and regulations.
The Power of Scrutiny
- Public pressure and scrutiny are crucial for driving change in tech companies' practices.
- IBM actively addressed bias after Gender Shades, while Amazon was less receptive initially.