Amazon, one of the leading purveyors of facial recognition technology (with their own Amazon Rekognition), has joined Microsoft in backing calls for a legislative framework to regulate the use of facial recognition technology.
In a recent bog post, Amazon’s AWS Vice President of Global Public Policy Michael Punke, re-iterated the benefits of using their facial recognition technology in law enforcement after reports of biases in the software, and outlined proposed guidelines for the responsible use of the software by private, commercial, and government entities alike.
Punke noted that Amazon has “talked to customers, researchers, academics, policymakers, and others” while putting together the proposed guidelines. The six guidelines outlined by Punke are:
- Facial recognition should comply with the law, including laws that protect civil rights
- Human review is necessary to ensure a prediction does not violate civil rights
- A 99% confidence score threshold is recommended
- Law enforcement should be transparent about how they use the technology
- Public should be notified when video surveillance and facial recognition are used together in public or commercial settings
These largely reflect the ethical framework that Microsoft put forward in December of 2018, including transparency, notice and consent, and fairness.
This is an initiative that we at ESOMAR welcome, as these are principles we have advocated in favor of for over 70 years and are enshrined in the ICC/ESOMAR International Code.
Where these ever-increasing pushes for ethical and legislative frameworks surrounding facial recognition take us remains to be seen, and we look forward to these same guidelines being applied to all types of personal data beyond just facial recognition!
