Why is Big Tech hellbent on making AI opt-out?

My annoyance with the AI additions is only slightly higher than the constant UI changes in, say, a Google UI. So f&@@&?@&!& annoying to login to do something and hav things moved around for zero benefit, but it really is the icing on the cake when they put a pop-up over the UI declaring the change with pride, which further distracts me from accomplishing the task I need to.

So now whenever I type in GMail I have an annoying "Polish" button to distract me from the writing. Not the worst change, but certainly a net negative. And why would I ever want to translate to Polish anyway? You won't even find any Polish text in my email anywhere.

Because such decisions are for shareholders and not users. I guess we should be thankful we can disable it at all in some cases.

Apple Intelligence was particularly a let down. It just got in the way, adding noise and distraction and not adding anything useful.

Stock prices and promotions.

If you're a product person at a Big Tech, making your AI feature opt-out means you can justify a promotion with "My new generative AI feature has seen record fast adoption with 200 MILLION Weekly Active Users"

And if you're an executive, you can claim any customer value comes from the AI features, which the stock market likes right now.

> Regulators worldwide are keen to ensure that marketing and similar services are opt-in [...] but for some reason, forcing AI on customers is acceptable. [...] there needs to be a moratorium on adding AI services without first asking customers for consent

Consent should be required if you're collecting private data for marketing or for training an LLM/etc.

But, I don't think adding some ML functionality to an application (say, OCR to a PDF reader, or translation to posts on a social network) inherently requires getting customers' consent any more than adding any other feature would.

I can appreciate the annoyance over "formerly pristine applications cluttered with AI features" since it's the current trend - but see it as similar to the annoyance over flat design (which rarely even had an opt-out) as opposed to something that needs a legal prohibition.

Because they believe (correctly, I think) that AI is going to be foundational to most of the computing experiences we have, moreso over time.

I don't think it's particularly absurd to make an analogy to other foundational technologies. When relational databases were invented, and enabled a ton of new & better capabilities across a host of user-facing software, it would have been strange to let users "opt-out" into the more limited, simpler experiences that could still be supported on earlier data architectures. And what would it even mean, practically? Maintain two separate data management systems, one relational and one legacy?

AI is almost certainly going to be like that, threaded throughout the designs of essentially everything above a certain level of computational complexity.

From a general consumer point of view? If it has AI,it's the latest and greatest. It's got to be better than something without AI.

It's opt-put because most people won't opt-out, if only because they can't be bothered. Lacking strong privacy laws, the user agreements probably say that all their data can be used for any variety of uses. "We're going to use your data to help make our AI features better? Oh, sure, that's great!" (Most people won't think of the privacy implications.)

There's an opt out?

Firstly, To justify easing the product into subscriptions if it isn't already. Secondly, for product teams to justify their work. Thirdly, to signal that new features are being developed. Finally, when generative AI works in a product is is pretty awesome (but sadly it doesn't work well all the time).

Have to justify all those billions spent on it somehow.

They want data / feedback to improve the model