It may be surprising that Microsoft has made repeated calls for lawmakers to regulate facial recognition, since it's a technology the vendor offers itself. But that's exactly what Microsoft chief legal counsel Brad Smith did yet again this week when he published another blog post asking regulators to step in.
Let's be clear: it's not like Microsoft wants to make it harder to further develop its Windows Hello or other offerings that rely on facial recognition. Quite the opposite. Smith notes that some of facial recognition's problems stem from its shortcomings. (How short? Well, a civil rights organization discovered that the UK's facial recognition is wrong 98 percent of the time, which means it's more likely to be actively harmful than actually helpful.) Microsoft has no intention of abandoning facial recognition, it simply isn't letting those interests obscure the technology's poses risks.
Microsoft spoke with "technologists, companies, civil society groups, academics and public officials around the world" about the issues posed by facial recognition after Smith's initial blog post. Those conversations led it to believe that governments need to address three problems with facial recognition: that current limitations can lead to biased outcomes that may run afoul of discrimination laws, that its spread could lead to invasions of privacy and that "the use of facial recognition technology by a government for mass surveillance can encroach on democratic freedoms."
Many of Smith's proposed solutions to these problems come back to requiring companies to be transparent about how facial recognition technologies are developed and used. He called for independent companies to have access to these technologies, for example, so they can determine if they're biased. He also suggested that stores using facial recognition technology to track customers--a tool being developed by Microsoft, Amazon and other companies--be required to inform shoppers of their use so they can decide whether or not they want to shop there.
Smith also called back to his claim that facial recognition's advance could make the year 2024 seem like 1984 if left unchecked.
"One vision of the future would require that citizens must evade government surveillance by finding their way secretly to a blackened room to tap in code with hand signals on each other’s arms--because otherwise cameras and microphones will capture and record their faces, voices and every word. Orwell sketched that vision nearly 70 years ago. Today technology makes that type of future possible. ... But not inevitable," he said.
The key, according to the exec, will be to "ensure that governmental use of facial recognition technology remain subject to the rule of law" via new legislation.
Not that all of the work should fall on regulators; Smith also published Microsoft's principles for developing facial recognition technology and said the company would go into further depth on them next week.