Weak public interest demand for AI

AI development rarely responds to demands from the public

Addressing the neglect and obstruction of public interest demand in the AI field is a priority for The Collaborative. This problem definition outlines our understanding of the problem and how we see it evolving. We expect this statement to evolve as we learn and in response to changes in the field.

If AI is to serve people and communities and not the other way around, its development must be driven by public interest outcomes: what are the specific problems that AI is trying to address and who defines them? Seen through this lens, the development of artificial intelligence has become divorced from public demands and needs.

AI can feel inevitable. For educators, learners, doctors, patients, for most of the public, AI is coming too fast, leaving little room for choice and agency, let alone time to imagine a different, alternative future where AI serves them. It is the demands of leading companies, not the public’s, that are defining the AI field.

Demand for choice and participation unmet

The public is increasingly making its voice heard, from local opposition to data center construction to coalitions concerned about AI’s impact on nature. There is pushback against decisions made without meaningful public participation. Examples include the use of AI systems to detect social assistance fraud and mistaken claims in the Netherlands and Australia, as well as to process visa applications to the UK.

These kinds of situations provoke a generalised loss of trust and agency. They can manifest either as concerns about corporate power and state surveillance, as concerns about cultural imposition and loss of national sovereignty, or as the fears of workers who have lost or fear losing their jobs to AI.

Seen through this lens, the development of artificial intelligence has become divorced from public demands and needs.

If you live as an average member of industrialized society, anywhere in the world, chances are you have little choice but to use artificial intelligence and are subject to decisions based on automated decision-making. Sometimes this is because AI is deployed to set prices, detect fraud, make decisions about insurability, or recommend products in systems that are invisible to the public. Other times, opting out is nominally possible but cumbersome and technically demanding. It has become difficult to compose an email or use a new mobile phone without being nudged to use AI.

Agency in the use of technology has become a privilege today. The privilege can mean avoiding loyalty programmes that require giving up privacy; having access to schools that are thoughtful about technology use; only working with doctors who protect patient data. Most individuals have less control over their privacy than companies who use AI at the enterprise scale. Properly aligned demand will make it possible for people around the world to use AI when it helps them accomplish things they want to do, and to refrain from using it when they choose to do so. This is what AI development driven by public interest demand would look like.

Learn more about the problems we are focused on: