Computer Vision Research Is a Surveillance Incubator

Perhaps it was always the height of naïveté to assume that forty years of research into “computer vision” – the field of artificial intelligence that allows computers to interpret images – would only make the world a better place. Now a landmark study published in Nature reports:

• “Our findings challenge narratives that most kinds of computer vision and data extraction are largely benign or harmless and that only a small portion is harmful. Rather, we found that the computer-vision papers and patents prioritize intrusive forms of data extraction.”

Let’s be clear about what that means.

• The study found that fully 86 percent of the field’s derived patents focused on using computer vision technology to create and extract human-related data, about bodies, body parts, and the spaces where such things are found. In other words, it is technology that tracks individuals and what we do.

The field of computer vision research, says ScienceBlog, has become “a vast network that transforms academic insights into tools for watching, profiling, and controlling human behavior.”

• Australian academic Jathan Sadowski told The Register that the discipline’s research conveniently meets the insatiable surveillance needs of its friends in high places – the military, law enforcement, and corporations: “Computer vision's focus on human data extraction does not merely coincide with these powerful interests, but rather is driven by them.”

The Nature study also discovered that, to help normalize their voyeurism, researchers in the computer vision field had adopted some clever linguistic tactics, including referring to humans and our body parts as “objects” – a kind of dehumanizing rhetoric that brings to mind the worst of the previous century. The obfuscating language, notes Sadowski, has the added bonus of abstracting “any potential issues related to critical inquiry, ethical responsibility, or political controversy.”

Such studies underscore a pressing reality, namely that we are rushing full tilt into uncharted territory without brakes and without any guardrails on the road. And in some cases, without any roads at all or even any maps.

Technology, and especially AI, doesn’t have to exist in the absence of privacy and accountability. With that goal in mind, one of the leaders of the Nature study, Dr. Abeba Birhane, recently established the AI Accountability Lab at Trinity College Dublin. She represents the best of us in the struggle against an unrestrained surveillance state. The whole site is worth a deep dive, but we’ll leave you with this excerpt:

• “AI systems are integrated hastily into numerous social sectors, most of the time without rigorous vetting. As a result, AI systems built on social, cultural, and historical data and operating within such a realm tend to diminish fundamental rights, and keep systems of authority and power intact.”

AI Accountability Lab reminds us that any new powerful technology inevitably has political and personal consequences. When technology drives the times, philosophy and prescience are needed as never before.
​
Congress and the states need to enact – use whatever metaphor you like –  legal maps, guardrails, brakes. In the end, they all represent the same idea – a return to the privacy guaranteed to all American citizens in the Bill of Rights.

DONATE & HELP US PROTECT YOUR PRIVACY RIGHTS