Algorithmic Mediation and Consumer Agency: Toward a Theory of Choice in AI-Driven Digital Markets

Artificial intelligence (AI) sits right at the intersection of consumers and markets these days. Algorithms run the show across e-commerce sites, search engines, streaming apps, social media, travel sites, financial tools, and even the latest crop of generative AI assistants. They filter, rank, recommend, personalize, predict—sometimes they even make choices for you. Sure, these tools cut down the avalanche of information and streamline transactions, but they also mess with how people figure out what they like and how they decide. This paper lays out a theory for how algorithms shape consumer agency, weaving together research from consumer behavior, recommender systems, behavioral economics, algorithmic governance, autonomy, and responsible AI. The main point: agency in digital markets isn’t just about picking from a bunch of options. Real agency depends on whether people can actually spot their true alternatives, understand their choices, stick with preferences they genuinely care about, resist algorithmic nudges, go beyond just following recommendations, and really influence the systems that craft their options. The paper breaks this down into four ways algorithms mediate things: they set up what options you see, guess what you like, steer your behavior, and gradually mold your preferences. From here, the paper introduces the idea of “bounded algorithmic agency†—meaning people still make deliberate choices, but always inside a digital world built and constantly tweaked by platforms and AI. The model says algorithmic systems open the door to efficiency—less time searching, sharper decisions—but also shrink agency by making people feel less in control, killing off chances to explore, and looping back into their preferences. It flags transparency, the power to challenge recommendations, system setup, preference matching, user skill, and the importance of the decision as core factors that shape these effects. The paper offers ten hypotheses for researchers to dig into and wraps up with this: AI-driven markets demand a fresh take on consumer choice—one that goes beyond the end decision and zeroes in on the systems deciding what’s visible, what counts, and what’s even possible to pick in the first place.