08-17-2026
Companies rushing to replace human workers with AI may be creating a problem they haven't fully considered: they may no longer own what they produce.
Amazon, Microsoft and dozens of other firms have announced layoffs tied to AI adoption, drawing ire from many consumers. But this substitution carries a hidden risk. Under current U.S. law, intellectual property (IP) created without meaningful human involvement cannot be patented or copyrighted. A pharmaceutical company that uses AI to develop a new vaccine, or a publisher that uses AI to write a book, may find itself unable to prevent competitors from copying the result.
This tension is at the heart of new research by Daniels School of Business alumnus Logan Emery and co-authors Robin Döttling and Shuo Zhao, examining how IP systems and AI shape firms' investment decisions.
IP rights have long driven business innovation. A patent or copyright gives firms a temporary monopoly to recoup development costs and fund future R&D. Without that protection, competitors can free-ride, eroding the incentive to innovate.
Generative AI is revolutionizing this process by dramatically lowering the cost of innovation. However, heavy AI use may render innovations ineligible for IP protection. Firms using AI in innovation must therefore balance these considerations.
One important decision still facing IP systems is how much to differentiate policy based on AI use. A sufficiently differentiated policy can make it difficult enough to obtain protection when using AI that it kills AI use in innovation altogether, so long as firms still value the monopoly protection the IP system grants.
But if AI lowers costs enough, firms may not need IP protection to justify innovation. AI could therefore kill the IP system. However, that is true only if human involvement carries no social value. If society values the joy of creativity, the dignity of work, or knowledge spillovers from keeping humans involved in innovation, then it may be worthwhile to maintain a differentiated policy. The IP system would then shift from subsidizing innovation to subsidizing human involvement.
While optimal IP policy is still being debated, consumers have already formed strong opinions about AI output, reflected in the popular term “AI slop.” Evidence shows consumers consistently prefer human-created art, music and books over AI-generated alternatives, even if they cannot tell the difference. This creates a classic adverse selection problem: firms relying heavily on AI can mimic those using human talent, driving down the premium consumers pay and, in the worst case, triggering market collapse.
The IP system can help. When consumers see that protection is tied to human involvement, they expect more human input and raise their willingness to pay, strengthening firms’ incentives to use human talent. The granting of protection can then also serve as a certification of human involvement if consumers observe the granting decision. And if AI-heavy firms still rely on IP protection to make innovation profitable, the IP system can eliminate AI slop and prevent market collapse.
The implications for copyright and patent policies differ substantially. Creative works like music, art and writing are more likely to carry social value from human involvement, and consumers consistently say they prefer these works made by people. That combination pushes copyright offices toward strict rules separating human from AI work, potentially eliminating AI use in creative fields.
For patentable inventions, however, quality likely outweighs the social value of, and consumer preference for, human involvement. As AI improves, firms that depend on patents may need that protection less, putting the patent system at risk of becoming obsolete.
Logan Emery is an Assistant Professor of Finance at the Rotterdam School of Management, Erasmus University Rotterdam. His research focuses on digitalization, investments, political connections, and innovation. He holds a PhD in Management (Finance) from Purdue University, completed in August 2021, where he wrote his thesis on “Essays on the Financial Implications of Web Traffic,” and an MS in Economics from Purdue University, awarded in May 2019.
