What Targeted Ads Can Teach You About Data Science Ethics and Privacy

What Targeted Ads Can Teach You About Data Science Ethics and Privacy

Have you ever looked at a pair of shoes online, decided not to buy them, and then seen ads for those same shoes everywhere you go?

It can feel like the internet is following you.

While it's easy to dismiss this as a minor annoyance, it raises a much bigger question: When does personalization become an invasion of privacy?

That's one of the questions explored in the University of Michigan's Data Science Ethics online course available on Michigan Online, where H.V. Jagadish uses targeted advertising to show how data can improve our digital experiences while also creating important ethical challenges.

Jagadish has had a distinguished career spanning more than three decades. He is currently a distinguished professor of electrical engineering and computer science in the College of Engineering at the University of Michigan.

Why do targeted ads work?

From a business perspective, targeted advertising is easy to understand.

Imagine a retailer showing an ad for running shoes to random people online. Most won't be interested.

Now imagine showing that same ad to someone who recently searched for running shoes or spent several minutes comparing different models. The chances of making a sale are much higher.

That's where data science comes in. By recognizing patterns in our online behavior, companies can recommend products that are more relevant to our interests.

When it's done well, everyone benefits. Businesses reach people who are more likely to be interested, and consumers see fewer irrelevant ads.

But there's a catch.

When personalization goes too far

In the online course, Jagadish compares online advertising to shopping in a physical store.

Suppose you walk into a shoe store and browse a few pairs. A salesperson might recommend similar options or answer your questions. If you decide not to buy anything, they usually let you continue shopping—or leave the store—without following you around all afternoon.

We understand that social boundary.

Many online advertising systems don't.

Instead, they continue showing the same products days—or even weeks—after you've lost interest. The technology is very good at recognizing interest. It's much less effective at recognizing when that interest has ended.

That difference highlights an important lesson in data science ethics: Just because data can be used doesn't mean it should be used forever.

The bigger privacy picture

Targeted advertising is only one example of how organizations use data.

Every day, we leave behind digital traces through activities like:

• Browsing websites
• Using mobile apps
• Joining loyalty programs
• Interacting with government services
• Using connected devices and smart sensors

Each of these interactions may seem insignificant on its own.

Combined, however, they can paint a remarkably detailed picture of who we are—our interests, routines, habits, preferences, and even aspects of our daily lives that we never intended to share.

Organizations known as data brokers collect information from many different sources, making it possible to draw conclusions that no single dataset could reveal on its own.

Even metadata—information such as when, where, or how long an activity occurred—can expose surprisingly sensitive patterns without revealing the actual content of your communications.

Designing data systems people can trust

The goal isn't to stop using data.

The challenge is learning how to use it responsibly.

That's why many organizations are adopting privacy by design, an approach that builds privacy protections into products and systems from the very beginning instead of adding them later.

For data scientists, ethical decision-making means asking more than whether an algorithm improves accuracy or increases revenue. It also means asking:

• Does this respect people's expectations?
• Are we collecting only the data we truly need?
• Could this use of data undermine trust?

These questions don't always have simple answers, but they're becoming increasingly important as data shapes more of our lives.

Go Beyond Targeted Ads: Learn to Navigate Data Ethics

Targeted advertising shows how quickly useful data can raise difficult questions about privacy. But it's only one of the ethical challenges facing people who collect, analyze, and make decisions with data in their professional role across industries and occupations.

In Michigan Online’s Data Science Ethics non-credit online course, you’ll examine a broader range of questions about privacy, consent, fairness, data ownership, and the consequences of data-driven decisions.

What You’ll Learn

• Examine the ethical and privacy implications of collecting and managing big data.
• Explore the broader impact of the data science field on modern society.
• Understand who owns data, how we value privacy, how to receive informed consent, and what it means to be fair.

The course takes you beyond the privacy questions introduced here. Through real-world cases, you’ll consider issues such as informed consent, data ownership, anonymity, algorithmic fairness, surveillance, and the societal consequences of data science—and build a framework for thinking through ethical questions when the answers aren't simple.

Enroll in Data Science Ethics to strengthen your ability to recognize ethical challenges and make more thoughtful decisions about how data is collected, analyzed, and used.