Ethical questions in AI
An AI future
Edwin proposed reading two papers concerned with the future blessings that AI
might bring us. They've got quite a large scope, but could provide some good
topics for discussion.
Privacy and pattern recognition
A slightly different discussion was proposed by Dick, on some current applications
of pattern recognition, image processing and AI techniques in areas where privacy
plays a large role. Perhaps the issues raised here are complementary to the more
far-reaching papers above.
The material consists of a number of small articles from newspapers etc. Not everything
is equally relevant.
Face recognition:
- Unmasking criminals,
a BBC Tomorrow's World story on face recognition in CCTV images of crimes being committed
- In general CCTV images:
Pedestrian tracking in car parks:
Monitoring using CCTV (closed-circuit television) in general:
Data mining:
AI software for blocking 'dirty pictures':
- Smut filter blocks all but smut
"It actually evaluates the page before it downloads (by) breaking apart the
images, breaking apart the pixels, looking at shadings and color,
common positions. We train the system. There are over 10 million neural
networks involved in the thing," Beecher says.
General privacy sites: