The UX Ethics Gap: Why Amazon’s AI Products Raise Red Flags

Are we in favor of Amazon’s design decisions?

Are we in favor of Amazon’s design decisions? There’s no doubt that we are all extremely reliant on Amazon as a delivery service. In fact, as I’m writing this, I had sworn to limit my Amazon spending and purchases in 2025… buuuut buying online and getting packages delivered directly home is so convenient that sometimes I break my own rules.

Amazon is also a powerhouse in terms of its streaming services, as well as its drop-shipping services and AWS services for small businesses and enterprises alike. Amazon is a giant, a corporate behemoth, too big for its own good.

And we as consumers walk a fine line between over-reliance on the convenience their services offer, while also being critical of the way that they run things….

How many times have we read negative news headlines about Amazon’s poor treatment of warehouse workers, delivery drivers, even under-cutting small businesses that sell via their e-commerce platform?

But let’s dig further into Amazon’s UX e-commerce practices, some might even be described as dark patterns that they utilize to sell aggressively to consumers:

• Their one click buy is probably the most famous. Amazon is the only company to have patented the one-click buy. Apple happens to license this capability for their Apple music store.

• Amazon defaults to subscribe when you try to purchase an item from the e-commerce platform, rather than just letting you make a one-time purchase right away.

• Amazon routinely asks you to upgrade to Prime for faster deliveries, making it very hard to say no, especially when you want a rush order.

Let’s also get into Amazon’s Alexa, and when it made news headlines because of putting in accidental orders after ‘hearing’ things in passing or from TVs in homes.

This brings us to AI design and designing for uncertainty. Alexa’s AI in this case worked by taking in inputs (i.e., overhearing our voices, noises, commands) and creating an output (i.e., retrieving information, ordering).

Every AI output has some amount of error rate- for example, an AI system can be 99% certain that the output is correct or it can be 40% certain that the output is correct. Depending on how it was designed, it will suggest the output anyway.

Clearly in Alexa’s case, Alexa executed on orders with low uncertainty rates (i.e., overhearing conversations or the TV) which only shows how Amazon leveraged dark patterns to keep aggressively selling via AI and Alexa. No surprises there.

But what was the cost of their dark patterns and low certainty outputs from Alexa’s AI? It cost them winning the AI race. Everybody stopped trusting Alexa, only using the device for basic tasks like checking the time, calling a friend or family member, etc.

And then ChatGPT came along… The clarity of ChatGPT outputs, giving us options when the AI was not so certain about our requests or inputs, is reassuring. It builds trust. And this is why ChatGPT is where it is today, while Amazon’s Alexa is left collecting dust in people’s homes. Yikes. Comparing ChatGPT to Alexa is like comparing night and day.

This is why ethical AI design is so incredibly important. Amazon leveraged dark patterns and short term sales goals, completely missing out on billions of dollars because they favored short term profits over long-term investment into an innovative AI product.

On the other hand, ChatGPT focused on rigorous training of their large language models (LLMs) with thoughtful, nuanced and complex responses (AI outputs) to user requests (inputs). And I’m not saying that ChatGPT is perfect! All products always have room for improvement, which is what makes design so exciting, a constant work in progress.

If you enjoyed this case study, consider taking our 0-1 AI Product Innovation course ($490) with over 6 hours of video content and case studies like this one. The course walks you through building an AI product or feature to add to your UX portfolio. You can register for the course for free to watch our 60-minute course preview.

If you would like to get a taste for what our online modules feel like, watch our free YouTube lesson on UX Building Blocks: Atomic Design, UX Laws & AI Design Made Simple. We touch on AI design and give you a sneak preview of content from our online 0-1 AI Product Innovation course.

As always, rooting for your success!

Samaya

P.S. We updated our VIP UX jobs board just yesterday with over 30+ brand new entry level UX designer, UX researcher, product designer, and UX intern roles. Subscribe to get access to our password-protected VIP jobs board, updated on a weekly basis. You can access our free UX jobs board here, updated monthly.

P.P.S. We are continuing to host live events and a Remote Women in UX Design conference in the month of May! Register for UX events and the conference before seats run out! In case you missed our past live UX events, subscribe to our YouTube to get access to UX lesson recordings.