This article has been translated with DeepL.
Critics can pave the way for new AI innovations
- Published: 20 Nov 2026,
- 8:59 AM
- Updated: 20 Nov 2026,
- 8:59 AM
Those who are most enthusiastic about new technology aren’t always the best at implementing it. A study of AI innovations in healthcare shows that skeptics can play a crucial role. But their rigorous testing also causes more innovations to fall by the wayside.
Before large organizations decide to adopt new technology, pilot tests are often conducted. This gives entrepreneurs with innovative solutions the opportunity to test their solutions in a real-world setting.
In a two-year ethnographic study, Rebecca Karp of Harvard Business School tracked how six startups’ AI innovations in diagnostics were tested in three different healthcare organizations. A total of 13 pilot tests were studied, during which she closely observed how internal committees—comprising doctors, nurses, and administrators, among others—determined how the new technology would be tested.
The setting determines what is tested
The study, published in the top-ranked academic journal *Administrative Science Quarterly*, shows that who participates in the decision-making process and their attitudes toward the technology influence what is tested and, consequently, the results.
The internal committees created for each pilot project guided the tests by, among other things, limiting their scope and determining how challenging they would be.
Same technology, but different decisions
One of the companies had developed AI technology to detect skin changes, and that solution was tested in two different healthcare organizations. In one of them, where a majority of the committee members were skeptical, advanced features that had not been tested before were evaluated. The skeptics hoped that the innovative AI solution would fail.
In the other organization, those who were positive about the technology were in the majority. There, the tests were limited to tasks that the healthcare staff were already able to perform relatively well.
In both cases, the test results were successful. However, only one of the organizations began using the technology. And that was the organization that had put the technology to the test through more advanced trials. In the other organization, the benefits were not considered significant enough to justify adopting the technology.
Successful tests aren’t enough
In the study as a whole, trials were more likely to succeed when proponents were in the majority. More innovations were also adopted in these cases. However, among the successful trials, the technology was more likely to be actually adopted when critics had guided the trial process.
Rebecca Karp also shows how the majority in the committees controlled the testing process, including by limiting what would be tested, assigning responsibilities, and influencing which issues were discussed. Outwardly, the tests appeared objective, but in practice, they were designed to support the majority’s view.
The study shows that it is not only the characteristics of the technology that determine whether an innovation begins to be used. How it is tested is also important. The paradox is that those who are in favor of adopting a new, innovative technology may actually limit the opportunities for its adoption, while critics can help create better conditions for the technology to be adopted.
About the article and the author
The article Crossing the Design–Use Divide: How Organizations Select Innovations to Adopt is published in the academic journal Administrative Science Quarterly.
The author is Rebecca Karp, Harvard Business School, USA.