Artificial intelligence may make individual research tasks faster while quietly weakening collaboration, according to an essay published by Research Agenda that asks leaders to account for the human contact lost when technology removes everyday friction.
The author describes the pattern as the “Waymo effect,” drawing on the appeal of a driverless ride that avoids conversation, negotiation and social obligation. The immediate benefits are easy to measure: privacy, convenience and uninterrupted time. The harder-to-see cost is the chance encounter or informal exchange that never takes place.
The essay applies that idea to research and education. AI systems can help a scientist search literature, summarise material, draft text or work through a problem without asking a colleague. Each use may be efficient on its own. Repeated across an institution, however, those choices may reduce the number of conversations through which researchers compare assumptions, encounter unfamiliar methods and form new collaborations.
This is not presented as an argument against AI or automation. The concern is that institutions often count the minutes saved by a tool while failing to track the relationships and ideas that disappear when people no longer need one another for routine assistance. Friction can be frustrating, but some forms of friction create contact. A question put to a colleague may produce an answer, a correction or an unexpected connection that a solitary workflow would not generate.
Research leaders therefore face a design problem rather than a simple decision about adopting or rejecting AI. If automated tools become the default route for every small task, organisations may need to create deliberate opportunities for exchange elsewhere. Meetings alone may not replace informal interaction, especially when tightly structured agendas leave little room for exploratory discussion.
The argument also challenges how productivity is evaluated. Faster completion of isolated tasks is visible, while a collaboration that never begins cannot easily be recorded. Institutions could see short-term gains in output without recognising a gradual decline in intellectual diversity or the movement of ideas between teams. That risk is particularly relevant in research, where important advances often combine knowledge from different disciplines.
The essay is a perspective rather than an empirical study, and its “Waymo effect” is a proposed concept, not a demonstrated causal law. Its value lies in identifying a question that conventional efficiency metrics overlook: whether reducing social effort changes the kind of knowledge an institution can produce.
As AI tools spread through laboratories, universities and publishing, the practical test will be whether organisations can preserve the convenience of individual assistance without turning research into a collection of isolated workflows. The author urges leaders to treat collaboration as infrastructure that requires active maintenance, not as an automatic by-product of placing talented people in the same organisation.
