Skip to Content Skip to Footer
A Call for Mutual Disclosure in AI-Driven Digital Marketing

A Call for Mutual Disclosure in AI-Driven Digital Marketing

Lawrence A. Crosby

illustration of hand pulling down curtain with rope

Lawrence A. Crosby on the uses of AI and machine learning in personalization, and the risks they could pose

It’s nearly impossible to ignore the attention given to artificial intelligence and machine learning in all fields—including business and specifically in marketing. The strength of AI/ML algorithms is their ability to make incredibly accurate predictions. Compared to conventional statistical tools and models, AI/ML can easily accommodate non-linearities and complex interactions while automatically learning from mistakes. With recent leaps in processing power, AI/ML is made-to-order for Big Data. The combination of Big Data and AI/ML has the potential to become a tremendous force for good in society, from predicting who will contract a disease and prescribing the best form of treatment, to forecasting the location of natural disasters associated with climate change.


AI is a bit of a catchall that refers to a variety of systems and programs designed to simulate human intelligence and decision-making—some simple and some sophisticated. Machine learning is a sub-category of AI that includes programs that can learn and adapt on their own, in response to trial-and-error experience and new data. The algorithm of neural networks fits here. Deep learning refers to what are (for now) the most advanced applications. These algorithms process data inputs into multiple layers of increasing abstraction (think neural networks on steroids). They are part of the capability behind voice and facial recognition.

Marketing may be ahead of the curve in some ways, especially when it comes to the application of Big Data and AI/ML to digital marketing. The promise is intense personalization based on a deep understanding of the individual’s needs, wants, preferences and habits, coupled with tailored, curated content that presents the right offer, in the right context, in the right channels and on the right device. This can create a win-win: greater sales and larger margins for the business and heightened satisfaction for the customer.

When it comes to customer satisfaction, it’s important not to put too much on the shoulders of the triple-play of Big Data, AI/ML and digital marketing. For one, digital marketing is generally associated with buyer search and is just one arrow in the marketing quiver. AI/ML can also address customer use and is already helping to power today’s smart products and the internet of things. Moreover, many have cited the important application of AI/ML to post-sale customer service (e.g., self-service, case classification and routing, chatbots, etc.).

Still, there’s at least anecdotal evidence that AI-driven digital marketing can enhance customer satisfaction if properly implemented. Various sources confirm heavy recruiting of AI talent by firms operating in the internet retail space. Firms (or parts of them) classified as internet retail by the American Customer Satisfaction Index received an average ACSI rating of 80 in 2019, compared to the national average of 76.4. Among internet retailers scoring 80 or higher on the ACSI were such companies as Amazon, Nike, Apple, Macy’s, HP Store and eBay, all active recruiters of AI talent. Other firms whose internet retailing efforts fared slightly worse on the ACSI—such as Walmart, with a score of 74—evidently see AI/ML as a path to improvement. As recently reported by CNBC, Walmart seeks to become a digital enterprise through AI applications in-store, direct and online.

But deep personalization is a lofty goal. Its pursuit, via Big Data, AI/ML and digital marketing, is fraught with some serious land mines. Close buyer-seller relationships are built on trust, which is about keeping promises and not engaging in opportunistic behavior. Customers expect that firms using their personal data in digital marketing efforts will stay true to their word regarding data protection (both stated and implied) and act in the customer’s interests. Trust violations can be difficult to overcome. Firms engaged in Big Data, AI/ML and digital marketing need to operate under a strong system of corporate governance that ultimately holds the board accountable for trust missteps.

But there’s also a growing body of literature that points to the risks of over-personalization. The criticisms are twofold: First, users are often slotted into narrow niches that become repetitive and don’t reflect the breadth of changes in their interests. This limits cross-selling opportunities and can lead to abandonment of the app, website or vendor relationship. The lesson for digital marketers is to mix it up.

The other criticism is the “creepiness factor.” This happens when a single digital interaction unleashes a cascade of retargeted content that often spans multiple devices. It’s especially disconcerting when that content deals with sensitive issues, such as health, divorce, finances or children. Digital marketers need to know their limits.

At a more strategic level, marketers seeking to pursue deep personalization via Big Data, AI/ML and digital marketing should consider theories of social behavior and relationship development, where the notion of disclosure looms large. For instance, social penetration theory posits that relationship closeness develops progressively through voluntary, mutual self-disclosures between the parties. A problem for many consumers is that their disclosures of personal information are one-way: I tell you lots of personal things about me and you’ll tell me what I can buy from you. This is not exactly mutual. A small step in the right direction is a privacy policy that discloses to the customer the type of personal information that’s being collected, how it’s used and with whom it’s shared.

Academics working in information systems have borrowed from the privacy and psychological stress literatures to suggest how companies can address customers’ privacy concerns around the involuntary disclosure of personal information that ends up in data repositories. Their suggestion is to offer customers a greater sense of control over the information, which may include enabling them to direct the sharing of personal information with third parties or allowing them to verify the accuracy of the information before sharing.

Personalization efforts involving Big Data, AI/ML and digital marketing will become more prevalent and sophisticated in the years to come. Perhaps there will be less focus on achieving hyperpersonalization and more on optimum personalization. Sure, many customers have no problem sharing their information in exchange for tailored messages to reduce the costs of search and make shopping a more pleasurable experience. But many others feel that the loss of privacy easily offsets those benefits. Hopefully, digital marketers will learn to use their tools in a judicious manner that respects these differences and provides safeguards to protect that most valuable intangible asset: the customer relationship.

Lawrence A. Crosby is the retired dean of the Drucker School of Management. He is the chief data scientist at the KH Moon Center for a Functioning Society, a part of the Drucker Institute at Claremont Graduate University.