I recently read a June 5, 2025 post by Ted Gioia titled The Force-Feeding of AI on an Unwilling Public (I read it thanks to my feed collection). The headline is on all fours with the post content and I am largely sympathetic to Mr. Gioia’s concerns about big tech companies forcing large language models into their products (I am also sympathetic to complaints about his very short paragraphs, but we will set those complaints aside). People may oppose AI appearing in previously AI-less products for a variety of reasons including, but not limited to, generally disliking such tools, the tools not being useful in the context of a specific product, data privacy concerns, or the lingering suspicion that the AI-enhanced product may soon become more expensive to accommodate new and unwanted features.
While I understand the concerns Mr. Gioia touches on in his essay, I understood them from afar. Much like how I managed to avoid learning what Temu is while it apparently saturated the internet with advertisements, I have mostly avoided having involuntary generative AI feeding. My most negative experience with generative AI has been the web scrapers assaulting our server, but that is neither here nor there. The disconnect between my own experience and Mr. Gioia’s inspired me to respond to one section of his post. I quote Mr. Gioia below:
What happens if we decide to boycott AI? The sad truth is that we can’t. It’s already too powerful.
“Can’t” is an decisive word choice. To Mr. Gioia’s credit, he illustrates what he means with examples. Referring to his own list of tech products, Mr. Gioia stated that in order to boycott AI, he would have to give up his email account, Google searches, the Microsoft Office suite, Amazon and other online retailers, Spotify and other music streamers which allow AI music, customer service at many companies, and more.
I agree with Mr. Gioia that some degree of interaction with AI is most likely unavoidable. “Contacting customer service” is a good example wherein the person seeking assistance may need help has no control over how it is delivered. Mr. Gioia ominously predicted that this may soon be the case with doctors offices, law offices, and job applications. I would be loath to deny the possibility.
Yet despite agreeing with Mr. Gioia that a perfect boycott of AI is most likely impractical if not impossible, I venture that his position, as presented, is unduly extreme. There may be cases where one has no choice but to use a product with unwanted and ever-encroaching AI features. Some products may be required for professional work. In other cases, there may be no non-AI infested alternative to a product. But for people who are irritated with AI finding its way into their computers and phones, I would recommend carefully considering the following:
- Do you need the offending product or service for your current use-case?
- If you need the offending product, are there ways to use it with intent to avoid the unwanted AI aspects?
- If you do not need the product or service, are there non-AI infested alternatives which may suffice for your use-case?
To illustrate what I mean, let us work with Mr. Gioia’s own tech examples.
- Email: There exist email providers which do not force AI on their customers. As of the publication of this article, I use mailbox-dot-org as my email provider and I interact with my emails almost exclusively through the Thunderbird desktop and mobile clients. People who do not like AI in their email providers and are not satisfied with the options to push it into the background may consider moving to an AI-free provided (admittedly easier if you own your email domain) or change how they interact with their email (e.g., using a local AI-free client such as Thunderbird).
- Google Search: I have written extensively about alternatives to Google Search. There exist alternative generalist and niche search engines and front-ends which either do not include AI summaries or make them optional. For those who need or prefer Google Search, notwithstanding its annoyances, I would recommend looking at end-arounds such as uBlock Origin filters or other Google-specific techniques. (I have previously discussed my preference for traditional web search over AI “search.”)
- Microsoft Office Suite: I draft most of my writing in markdown. When I need a word processor (or other software found in Microsoft Office), I use LibreOffice, which is free of cost and free and open source (LibreOffice Writer will soon be able to open markdown files natively). Microsoft Office is required in many workplaces and I understand that some people require it for their personal use. With that being said, Libre Office works perfectly for my use case and I suspect that many people could use it as a replacement for Microsoft Office. For those who have to use Microsoft Office, I would recommend carefully studying toggles which may moderate AI or other unwanted features.
- Amazon: I have encountered AI on Amazon in two areas. First, it is now impossible to search reviews directly – we must use Amazon’s slow and mostly-useless “Rufus” bot instead (aside: Rufus is a good AI helper name, but not a good AI helper). I also dealt with something that may have been AI while trying to reach a customer service representative via chat. I know other people have had problems with AI-generated digital merchandise on Amazon. Amazon seems intent on making AI unavoidable, but it is still possible to mitigate many issues by being an active searcher and paying attention to important information such as who is selling a particular product. It may also be worth considering whether Amazon is the best option for every purchase. My discussion of DRM-free games on Steam and Steam vs GOG may offer some ideas on how to think about Amazon.
- Spotify: I have never used Spotify or any other music streaming service. From having read other posts by Mr. Gioia, I suspect his concern with Spotify here is AI-generated music. While I am not competent to offer music streaming recommendations, I suppose the first thing to do would be to consider whether you need a music streaming service at all. If so, one may consider whether Spotify, given its apparent openness to cheaply-generated AI music, is the best option. If it is the best option, then I suppose that one should consider how to use it actively, that is intentionally choosing what to listen to or at least what types of music to listen to, and minimizing reliance (to the greatest extent practicable) on Spotify’s AI-enjoying algorithms. My essay on using feed readers may indirectly offer some ideas on how to think about listening to music.
Mr. Gioia is almost certainly correct that most people will not be able to “boycott” AI. But total avoidance is not necessary (or necessarily desirable, allowing for the possibility that one may oppose certain applications of large language models and other forms of AI while welcoming others in specific contexts). Instead, for people who are so inclined, I recommend considering in the case of offending products and services whether they are needed, whether their problems can be mitigated, and whether there exist alternatives. The answers in some cases may be more positive than others, but I suspect most people who have as many complaints as Mr. Gioia will discover that they can tidy up their digital environments to some extent or another.
(PS: Given how much Mr. Gioia loves Substack, I hope for his sake that Substack does not force AI on him. That would make for a very sad future post.)
