The next e-commerce battle is over algorithmic trust

Artificial intelligence (AI) has already started influencing how consumers search, compare and buy products through recommendation engines, chatbots, virtual assistants and personalized shopping tools. However, new research finds that human-like AI features alone are not enough to drive consumer purchase intention unless shoppers also see algorithmic systems as transparent and fair.

The study, titled "The Role of Algorithmic Anthropomorphism, Transparency, and Fairness in Shaping Consumer Purchase Intentions in E-Commerce: Evidence from Türkiye," was published in the Journal of Theoretical and Applied Electronic Commerce Research. Using Türkiye as an example of an emerging digital market, the study surveyed 384 online consumers and found that anthropomorphic AI can increase buying intention, but most of its effect is carried through consumer perceptions of algorithmic transparency and fairness.

AI shopping tools are changing how consumers judge online platforms

Online platforms now rely heavily on AI systems that recommend products, rank search results, offer chatbot support, personalize promotions and guide customers through purchase decisions. These systems can make shopping faster and more relevant, but they also create a trust problem because consumers often do not know how recommendations are generated or why certain products are shown.

The study examines three algorithmic design factors that are increasingly central to digital commerce: anthropomorphism, transparency and fairness. Algorithmic anthropomorphism refers to the extent to which an AI system appears human-like, natural, socially responsive or lifelike. In a shopping context, this may include chatbots that use conversational language, virtual assistants that respond with empathy or recommendation systems that seem to understand consumer preferences.

The findings show that human-like AI features can raise purchase intention. Consumers are more willing to buy when an algorithmic system feels approachable and easier to interact with. This matters for global e-commerce platforms because human-like design can reduce psychological distance between the shopper and the machine, especially when consumers are uncertain about digital tools.

The study also shows that anthropomorphism is not a stand-alone solution. Human-like cues work because they influence how consumers judge the system behind the interface. When AI feels more human, consumers are more likely to perceive the algorithm as transparent and fair. These perceptions then shape whether they are willing to buy.

This creates both an opportunity and a risk for online retailers. A chatbot that sounds helpful can build confidence, but it can also create a false sense of openness if the platform does not actually explain how recommendations are made. Human-like AI may make a system feel transparent, but perceived transparency is not the same as real disclosure.

Making AI more conversational may improve engagement for e-commerce firms, but it should be paired with clear and honest information about how the algorithm works. Consumers need to know why a product is recommended, what data are being used and whether the result is organic, personalized or commercially promoted.

Transparency and fairness carry the real commercial weight

The strongest finding is that most of the effect of human-like AI on purchase intention is indirect. The study tested a chain in which anthropomorphic design increases perceived transparency, transparency increases perceived fairness and fairness increases purchase intention.

Transparency had a positive effect on both fairness and purchase intention. Consumers who felt they understood how platform algorithms worked were more likely to judge the process as fair and more likely to consider buying. This finding speaks directly to the growing debate over black-box AI in commerce. Shoppers may accept personalization, but they are more likely to trust it when the platform explains the logic behind it.

Fairness also played a major role. Consumers were more willing to purchase when they believed that algorithmic recommendations and outcomes were fair, unbiased and justifiable. In e-commerce, fairness can mean that product rankings are not secretly manipulated, that sponsored products are clearly identified and that recommendations are not designed only to maximize platform profit at the consumer's expense.

The study found that roughly 69% of the impact of anthropomorphism on purchase intention moved through transparency and fairness. That means the business value of human-like AI depends less on friendly design alone and more on whether that design leads consumers to believe the system is understandable and fair.

This finding is important for both mature and emerging e-commerce markets. In countries where digital shopping is still expanding rapidly, consumers may be open to AI-powered services but also cautious about opaque systems. Türkiye provides one example of this broader pattern. Its online consumers are active users of e-commerce, but their responses to AI-driven platforms still depend on whether the systems feel fair, understandable and easy to use.

The finding also challenges retailers that treat AI design mainly as a conversion tool. Human-like chatbots and recommendation engines may increase engagement, but if consumers later suspect bias, manipulation or hidden commercial influence, those same tools may undermine trust. The more AI becomes part of the shopping journey, the more important it becomes for firms to show accountability.

For platform managers, practical steps are straightforward. Recommendation pages can include brief explanations of why a product appears. Sponsored items should be clearly separated from organic recommendations. Chatbots can explain when they are using browsing behavior, prior purchases or stated preferences to suggest products. Platforms can also give users more control over personalization settings.

Transparency should not mean overwhelming consumers with technical detail. The study's broader message is that consumers need meaningful clarity, not complex explanations. Simple, visible and plain-language disclosure may be more effective than dense technical descriptions.

Ease of use decides when human-like AI matters most

The study also examined technology acceptance, separating perceived ease of use from perceived usefulness. This produced a key finding for e-commerce firms designing AI systems for different user groups. Perceived usefulness directly increased purchase intention. Consumers who saw an e-commerce platform as useful were more likely to buy. But usefulness did not change the strength of the relationship between human-like AI and purchase intention.

The effect of anthropomorphic AI was strongest among consumers who found the platform less easy to use. For these users, human-like design appears to reduce friction, provide reassurance and make the shopping process feel more manageable. A conversational assistant can help them navigate uncertainty, understand recommendations and move toward a purchase.

Among consumers who already found the platform easy to use, human-like AI still had an effect, but it was weaker. These consumers may rely more on functionality, speed, search quality, recommendation accuracy and product relevance. They do not need as much reassurance from human-like cues because they already feel comfortable using the platform.

According to the study, e-commerce companies should not apply anthropomorphic AI in the same way to every customer. Human-like chatbots and guided support may be most valuable for users who struggle with navigation, need help understanding product choices or show signs of hesitation during checkout. Experienced users may respond better to faster search, stronger filters, better product comparison tools and clearer delivery or pricing information.

Personalization should apply not only to products but also to interface design. Platforms can use behavioral signals such as repeated searches, long decision time, abandoned carts or navigation errors to identify users who may need more conversational AI support.

Firms must also avoid using human-like AI to pressure uncertain consumers. If anthropomorphic design is used to push purchases without transparency and fairness, it risks crossing into manipulation. Ethical AI design should help consumers make informed choices rather than exploit confusion.

It is important to note that the study measured purchase intention rather than actual buying behavior. Its sample was mostly young, educated and based in Türkiye, so future studies across other countries and age groups would strengthen the evidence. The research also captured consumer views at one point in time, making long-term behavior harder to assess.