Artificial intelligence projects often fail for reasons that have little to do with the technology itself. According to a new study by Harvard Business School, employees frequently resist AI because they perceive it as a threat to their professional identity rather than because it delivers poor results.
Researchers Das Narayandas and Shunyuan Zhang argue that AI belongs to a category of “self-disruptive technologies”—tools that improve performance while making users feel less valuable, less influential or less recognized in their roles.
Identity, Not Technology, Drives Resistance
According to the Harvard Business School study, at least 30% of generative AI projects are expected to be abandoned, often because employees quietly resist adopting them rather than because the technology underperforms.
The researchers say previous explanations have focused mainly on organizational readiness or technical challenges, overlooking a critical factor: employees’ concerns about losing expertise, status and influence within their jobs.
As generative AI increasingly performs tasks such as writing, analysis and decision-making, many workers feel that the most meaningful aspects of their roles are being transferred to machines.
Three Ways AI Can Threaten Professional Roles
The study identifies three main ways AI affects employees’ sense of identity.
Role compression occurs when AI automates higher-value work, leaving employees with fewer responsibilities and tasks perceived as less meaningful.
Control shift reduces managers’ decision-making authority as algorithms increasingly guide business choices.
Span erosion weakens managers’ influence as more employees rely directly on AI rather than human expertise.
Rather than openly opposing AI, many employees adopt what the researchers describe as symbolic adoption—appearing to use the technology while limiting its real impact inside the organization.
Adoption Requires More Than Better Technology
According to the study, companies can improve AI adoption by demonstrating that the technology strengthens, rather than replaces, employees’ expertise.
The researchers recommend redesigning roles, maintaining human oversight in AI-assisted decisions, using AI to support rather than replace judgment, investing in employee retraining and ensuring visible executive support throughout the transition.
Instead of focusing solely on return on investment, organizations should also address employees’ concerns about professional identity.
As Narayandas concludes, companies are more likely to succeed when they show workers not only how AI benefits the business, but also how it enhances their own future roles and careers.
Photo: Vectorarte/ magnific.com
UK
IT
ES
DE
FR
RO
PT
HU
