New research has underlined the surprising extent to which pervasive surveillance of people and their habits is powered by computer vision research—and shone a spotlight on how vulnerable individuals and communities are at risk.

Analyses of over 40,000 documents, computer vision (CV) papers and downstream patents spanning four decades has shown a five-fold increase in the number of computer vision papers linked to downstream surveillance patents. The work also highlights the rise of obfuscating language that is used to normalize and even hide the existence of surveillance.

The research, conducted by Dr. Abeba Birhane and collaborators from Stanford University, Carnegie Mellon University, University of Washington and the Allen Institute for Artificial Intelligence (AI2), has been published in Nature.

Dr. Birhane directs the AI Accountability Lab (AIAL) in the ADAPT Research Ireland Center at the School of Computer Science and Statistics in Trinity College Dublin.

Dr. Birhane said, “This work provides a detailed, systematic understanding of the field of computer vision research and presents a concrete empirical account that reveals the pathway from such research to surveillance and the extent of this surveillance.”

“While the general narrative is that only a small portion of computer vision research is harmful, what we found instead is pervasive and normalized surveillance.”

Among the key findings were that:

1. The field has evolved linguistically, with a trajectory away from generic papers in the 1990s towards a heightened focus on analyzing semantic categories and humans and their behaviors in the 2010s
2. Surveillance has been increasingly hidden by jargon and obfuscating language that distracts from people being at the heart of the surveillance
3. Rights to privacy and certain freedoms are under threat

Additionally, the work indicated the top institutions producing the most surveillance are: Microsoft; Carnegie Mellon University; MIT; University of Illinois Urbana-Champaign; Chinese University of Hong Kong, and the top nations are: US; China; UK.

Dr. Birhane added, “Linguistically, the field has increasingly adapted to obfuscate the existence and extent of surveillance. One such example is how the word ‘object’ has been normalized as an umbrella term which is often synonymous with ‘people.'”

“The most troublesome implications of this are that it is harder and harder to opt out, disconnect, or to ‘just be,’ and that tech and applications that come from this surveillance are often used to access, monetize, coerce, and control individuals and communities at the margins of society.

“Due to pervasive and intensive data gathering and surveillance, our rights to privacy and related freedoms of movement, speech and expression are under significant threat.”

However, the researchers stress that a major, more hopeful takeaway from the work is that nothing is written in stone, and that this large-scale, systematic study can aid regulators and policy makers in addressing some of the issues.

Dr. Birhane said, “We hope these findings will equip activists and grassroots communities with the empirical evidence they need to demand change, and to help transform systems and societies in a more rights-respecting direction.”

“CV researchers could also adopt a more critical approach, exercise the right to conscientious objection, collectively protest and cancel surveillance projects, and change their focus to study ethical dimensions of the field, educate the public, or put forward informed advocacy.”