Whose Innovation Is It? Bidirectional AI Sensemaking and Reconfigured Trust in an Engineering Internship Mentoring Program
Plain Language Summary
Research on how workplaces adopt new technology usually imagines a senior person who evaluates an innovation and passes it down the hierarchy. In interviews with engineering interns and their alumni mentors, we found something closer to the opposite: interns were writing their companies' AI-use rules and being treated as the resident experts. This paper asks what that reversal means for how we think about technology adoption, and for whom organizations are willing to trust.
Contribution
We press on two assumptions in Diffusion of Innovation research, that the "change agent" is necessarily a positional leader and that interns are peripheral, low-trust participants, by offering bidirectional, if unpaired, accounts of generative AI sensemaking from both interns and mentors within the same program.
Research Questions
- How do interns and mentors make sense of generative AI within engineering internship workplaces, and how are AI-use norms and trust negotiated between them?
Methods
We analyze mid-summer, semi-structured interviews with 10 mentees (rising sophomores through seniors across Mechanical, Chemical, Electrical, Bio-, and AI-focused engineering) and 7 industry mentors in the Summer Engineering Experience, an alumni-run internship placement and mentoring program whose host companies are predominantly early-stage startups and small firms. Because the program draws on a small, identifiable pool, we generalize host-company identities, job titles, and other structurally identifying details, which also prevents formally pairing mentor and mentee transcripts as dyads; we coded inductively across the combined corpus using thematic analysis, with a second pass attending to the convergence of accounts from each side.
Implications
We argue for a more socially and relationally distributed model of technology diffusion in engineering workplaces, one that does not presume the change agent to be a positional leader. The evidence also raises open questions about what organizations are choosing to trust their most junior members to decide, particularly where interns hold governance-level judgment over sensitive and confidential information.