Amirhossein (Miros) ZohrehvandStrategy & AI

Research

My research studies AI in organizations and AI in research: what happens when AI takes on organizational roles, and how researchers can build methods and workflows on AI.

AI in Organizations

What happens when AI takes on organizational roles, such as in communication, coordination, and invention?

The Wade test asks whether generative AI can communicate as a specific CEO, and finds that believing a message is AI-written lowers how helpful people judge it, whatever its true source. Two projects take up coordination: one on generative AI and coordination at work, and a paired experiment that tests hierarchy in LLM agent teams by varying a manager’s authority to reject output and require revision. Inventions without inventors follows what firms disclose when human inventorship becomes uncertain. Across these the question resolves into three parts: whether AI can perform the role, whether people accept it there, and how credit and responsibility are allocated once contributions are hard to separate.

AI in Research

How can researchers build methods and workflows on AI, and verify what those produce?

The stream began with research methods: generalizing event studies with synthetic controls (Long Range Planning, 2024) builds credible counterfactuals for strategic events that happen too rarely for conventional designs. The AI projects keep that concern with valid measurement and credible inference. Measuring concept content from LLM activations tests whether a model’s internal representations yield valid measures of a construct the researcher defines, using ESG text as the test case. Recipes for creativity shows that evaluator design is a first-order choice in iterative generation, not a detail. TRACE extends the question to the research process itself, replacing unverifiable AI-use statements with workflows that are modular, traceable, and reviewable, derived from the practices computer science already uses to govern shared work. Much of this stream is published in computer science venues.

Earlier Foundations

My earlier work studied CEO social media and strategic leadership, and, in two computer science papers, AI avatars and user motivation and the social-media discourse around data breaches (both ICDSAI 2024). It is the foundation for the current agenda: it shows how a technology reshapes the relationship between leaders, audiences, and feedback before AI began occupying the role.

Selected Work in Progress

Journal Publications

Conference Proceedings

Computer Science

Management