E&Y sees troubling ethics priority gap between execs and pols on facial recognition
Biometric facial recognition, viewed through an ethics lens, looks markedly different to policy makers and business executives, according to a recent global survey. The perception gap could increase deployment challenges.
The report by professional services firm Ernst & Young examines how trust in AI between government and business leaders is not in sync.
Facial recognition is held out as an example of the situation.
Survey takers — 71 policy makers and executives from 284 companies across 55 countries — were asked to rank the importance of 11 ethical principles when it came to using facial recognition for service check-in.
Fairness and avoiding bias, and privacy and data rights dramatically outranked all other considerations for policy makers. Executives had no similar focus or priority.
Fairness and anti-bias ranked highest for government leaders, earning a survey score of 3.14. Privacy and data rights followed with a score of 3.09.
By contrast, ethics in innovation eked out a score of 0.14 among government leaders. The average of all principle scores was 1.36.
It is striking to note that the average score among business leaders was statistically the same.
The difference was that, without a dramatic result on ethics priorities, scores landed in a much narrower band.
No ethical principle scored higher than 1.88, and that was privacy and data rights. The next highest score was 1.62, for safety and security. The lowest score was a tie: 1.02 for explainability and well-being.
Executives prized six principles more highly than did their public service counterparts, which no doubt is reassuring to some who worry that facial recognition technology is a runaway freight train.
Those six were data access, explainability, human autonomy and agency, innovation, precaution, safety and security.
But Ernst & Young’s consultants made the point that clear priorities are needed with ethics, a business consideration that impacts all other aspects of product development, manufacturing, marketing and deployment.
best practices | biometric identification | biometrics | data protection | ethics | facial recognition | legislation | privacy | regulation