Why Executives Need to Take a Hard Look at AI-Driven Marketing
Artificial intelligence isn’t just working behind the scenes anymore. Today, it’s front and center—chatting with customers, answering their questions, guiding their choices, and even suggesting what they might want to buy. To many people, these interactions feel personal, seamless, and genuinely helpful.
However, new research suggests that AI recommendations have a bigger impact on what people buy than many business leaders might think. When these systems mix advice with advertising—especially if they don’t make it clear—they can cross ethical lines that go well beyond the usual marketing rules.
The question isn’t simply whether chatbots can recommend products.
The question is whether they should—and under what governance.
Why AI Recommendations Raise New Ethical Stakes
Traditional advertising—even when it feels pushy—works because everyone knows what an ad looks like. People can spot when they’re being persuaded.
AI makes that line blurry.
Consumers Over-Trust Conversational AI
A 2024 study found that when chatbots sound friendly and human, people trust them more—even if they’ve made mistakes before (Cheng & Liu, 2024). When a system “sounds helpful,” people assume it knows what it’s doing and isn’t biased, which makes them more likely to follow its advice.
Transparency Changes Outcomes
When people find out a recommendation comes from AI instead of a human expert, they’re less likely to trust it or want to buy (Svensson, 2024). This gives companies a reason to hide the truth—which is exactly why they need their own rules before outside regulators get involved.
AI Recommendations Are Often Biased
Recommender systems tend to amplify popularity bias, majority-user behavior, and systemic inequality. Recent work shows persistent issues such as miscalibration, stereotype reinforcement, and reduced option diversity (Ekstrand et al., 2023; Zhang et al., 2024).
If your AI always pushes partner or popular products but acts like it’s being neutral, that’s not really helpful—it’s bias dressed up as advice.
AI Personalization Can Cross Ethical Lines
Super-personalized recommendations use all kinds of user data—like preferences, worries, past behavior, and even mood—to make suggestions that are hard to resist. A 2024 survey of over 140 AI recommendation studies warns that these systems can take away some of our freedom to choose and push people toward decisions that help the company more than the customer (Zhang et al., 2024).
That’s why questions about AI ethics aren’t just for the tech team—they’re a job for company leaders.
Executive Checklist Sidebar: A Practical Checklist: Keeping AI Product Recommendations Ethical
-
Be Upfront About Paid Recommendations
- Are users told clearly if a recommendation is paid for, sponsored, or comes from a partner?
- Are these disclosures easy to spot in the chat, instead of hidden in fine print?
-
Keep Ads and Advice Separate
- Is the system’s logic for ads kept totally separate from the logic for genuine advice?
- Can your team walk you through exactly how and where this split happens?
-
Make AI Decisions Explainable
- Can you clearly explain why the AI suggested a particular recommendation?
- Is there a straightforward record of how the AI reached its decision?
-
Check for Bias and Undue Influence
- Have we checked if the recommendations are fair to everyone, regardless of demographic?
- Have we looked for signs of overly pushy recommendations or those that might take advantage of users?
-
Know Your Vendors and Models
- Do our AI vendors or models automatically favor commercial results?
- Are we making sure our partners are also being transparent with users?
-
Give Consumers Real Choices
- Can customers easily ask for suggestions that aren’t influenced by money or partnerships?
- Is this option clearly visible and easy to choose?
-
Build in Safety Checks
- Do we limit recommendations in sensitive areas like health, money, legal advice, or kids’ products?
- Do we have human reviewers involved where it really matters?
-
Prepare for Mistakes and Emergencies
- If a chatbot recommendation causes real harm—like financial, health, or safety issues—do we have a clear plan for what to do next?
- Who is ultimately responsible for fixing the problem and informing affected users?
If you can’t honestly check every box on this list, your chatbot isn’t ready to make recommendations—at least not in a way that’s ethical or trustworthy.
Where Executives Must Step In: Governance, Disclosure, and Control
AI-driven recommendations are where marketing, tech, law, and ethics all meet. That usually means no one really owns the problem—so everyone needs to care about it.
Executives should insist on the following:
Clear Disclosure of Monetized Recommendations
If sponsorship, affiliate deals, or partner arrangements influence ranking or recommendations, these must be disclosed directly in the chat interface. Research shows transparency matters (Svensson, 2024).
Separation of “Advice Logic” and “Ad Logic.”
If your chatbot is both giving advice and making sales, people should always know which hat it’s wearing.
Auditable Explanation of Recommendation Pathways
Executives must be able to answer:
“Why did our AI recommend this product at this moment to this customer?”
If your team can’t answer that question, your AI isn’t ready for real-world use.
Regular Bias and Manipulation Audits
We already know about echo chambers and popularity bias (Ekstrand et al., 2023). Any system that nudges vulnerable people—like kids, seniors, or families under financial stress—needs to be fixed or limited right away.
Human Oversight in Sensitive Domains
When it comes to money, health, legal advice, or big purchases like cars or insurance, AI should never make the final call on its own.
The Bottom Line: Trust Is the Real Product
AI recommendation systems aren’t just background tools—they actively shape what people choose. Their power to persuade is strong, subtle, and proven. If companies let these systems make suggestions without being open about it, they risk losing the one thing every brand needs most: trust.
The companies that will come out ahead in ethical AI won’t be the ones with the flashiest sales tactics—they’ll be the ones who combine new tech with real integrity, building AI that’s powerful, open, and responsible.
Leaders who take this seriously now will set the standards that everyone else will eventually have to adopt.
References
Baker, L., & Kim, S. (2025). From clicks to conversions: How AI shapes consumer trust, experience, and online buying behaviour. Advances in Consumer Research Journal. https://acr-journal.com/article/from-clicks-to-conversions-how-ai-shapes-consumer-trust-experience-and-online-buying-behaviour-1612/
Cheng, Y., & Liu, F. (2024). Social cues in AI chatbots and their effects on sustained trust after service failures. Humanities and Social Sciences Communications, 11(1). https://www.nature.com/articles/s41599-024-03879-5
Ekstrand, M. D., Zhang, Y., & Willemsen, M. (2023). System-induced effects in recommendation: A taxonomy and survey. arXiv preprint arXiv:2312.17443. https://arxiv.org/abs/2312.17443
Haque, A., Rahman, M., & Islam, R. (2025). The impact of AI-powered recommendations on online purchase decisions. International Journal of Business Research, 18(2). https://www.researchgate.net/publication/393716001
Svensson, E. (2024). Consumer trust and purchase intention when recommendations are labeled as AI-generated. Lund University Student Papers. https://lup.lub.lu.se/student-papers/record/9205608
Zhang, X., Li, H., & Kumar, A. (2024). A systematic review of AI recommender system impacts on consumer autonomy and preference diversity. arXiv preprint arXiv:2407.01630. https://arxiv.org/abs/2407.01630
