There's only one other major service that you pay for which also delivers ads to you. And that's cable television. You pay $50 or $60 or $80 or whatever a month for your cable connection plus channel packs. And the content is then delivered with a 5 minute ad break every 10 minutes. Why is this tolerated? Because TV before cable relied on a broadcast model, where anyone with a TV could view the content. So the *only* method to make revenue was to use ads. People got so used to ads in TV, it became such a core part of the medium, that when cable became dominant over broadcast TV, most people didn't think about it.
But if you're going to ask me to pay for your service, I refuse to watch ads. The reason ads exist is to make revenue off a product that the consumer does not pay for. If the consumer paid for the product, then you're already getting revenue and don't need advertising. Putting ads into a paid service is purely greed.
So what do I do? I exercise my rights as a consumer to go elsewhere. I don't have cable TV; I use Netflix with AdBlock turned on. I pay for the service, so I refuse to view ads.
Now, if you deliver content, then using a customer's viewing habits to recommend other content that the customer might like is reasonable. Netflix recommends shows and movies I might like. If the Xbox used your gaming habits to recommend other games you might like, based on your preferences, then that would be totally cool. Awesome in fact. But unfortunately they're not targeting game recommendations, they're targeting ads for products that have little or nothing to do with the service. And they don't target ads based on your actual wants or needs, they target ads based on A. who pays them the most, and B. demographic statistics. Thus what gets advertised is not tailored to what you want to buy, but tailored to whichever mass-marketed, generic thing the average [female genital cleaning product] in your age bracket might buy.
It would help if targeted advertising algorithms were more sophisticated, but they're not. Most algorithms are totally unable to differentiate between purchase-prone and non-purchase-prone activity. For example, I frequently view power supplies on Newegg and Amazon to keep up on pricing and features, and to look up specifications. I have no intention of buying a power supply; I have so many PSUs and am bombarded with so many more that I will likely never purchase another computer power supply in my life. However, since I view power supplies on Newegg and Amazon, these websites bombard me with advertisements for these products. Many targeted ad algorithms have similar flaws; they can't differentiate between genuine interest and intent to buy a product, and interest that will not lead to a purchase. It's a false positive. Thus users are frequently bombarded with advertisements that annoy them, and do zero good for the advertiser. This is the type of ad exposure, in fact, that breed resentment toward a product.
But it gets worse. Microsoft wants to put a camera into your living room to watch your face as you watch TV; to watch you interact with your friends and family; to listen to your conversations. And it then wants to aggregate that data to target ads? Every time you mention a product will be another data point for their algorithm to draw on. Do you figure their system is smart enough to avoid false positives? I suspect that consumers will be inundated with inane ads that are mistargeted based on an over-accumulation of data, a wealth of minutia feeding into the marketing myth that an algorithm can target ads with arbitrary precision, as long as it's fed enough information. What imbecilic, over-paid advertisers don't realize is that demographics and personal data become less relevant the smaller your sample size.
Basically, the Xbox One is bad for customers, it's bad for advertisers, and it's bad for Microsoft.