Ads Suck

My never-ending stream of pergola ads on Facebook.

The other day, I had the idea to roast some spicy cashews. So I did what most home cooks do: I searched for a recipe on my phone. What I got was a page on RecipesfromaPantry.com that was so laden with ads I could barely find the actual recipe. I was just trying to roast a nut, not book a Disney Cruise, buy cosmetics on Sephora, or treat plaque psoriasis with Otezla.

We’ve all found ourselves on pages like this. Innocently looking for some piece of information or trying to answer some question only to be inundated with pop-ups, auto-playing videos, full-screen interstitials, and un-closable banner ads. Often, these ad units are calling so many back-end ad servers that the entire site bogs down and you find yourself waiting like an idiot for the page just to load.

Then, heaven forbid you make the mistake of actually clicking on something. I clicked on a picture of a pergola on Facebook recently because I thought it looked cool. Now my Facebook feed is overrun with pergola ads—Pergolux, Bon Pergola, Outdoor Elements, Nordic Pergola, and on and on. It’s a veritable pergola-palooza. My Gmail is also, of course, populated with pergola promotions, since I’ve now been cookied as a pergola prospect. It’s ridiculous.

Before I spiral into a curmudgeonly gripefest about advertising, let me get to the point: the internet as we know it is optimized for advertisers. With AI, that may all be about to change . . . and possibly get worse.

The foundational quid pro quo of the internet—just like newspapers, magazines, radio, and television before it—is that you as the consumer get your media for free (or at least heavily subsidized) in exchange for enduring advertisements. Over most of the history of media, this proved to be a reasonable trade-off. Consumers got the information and entertainment they valued, and advertisers had a largely cost-effective way of reaching potential buyers. The advertising load we were each subjected to was manageable, even helpful in finding goods and services.

The internet changed that.

Today, the average American is exposed to an estimated 4,000 to 10,000 ads every day (there are eight just on this Forbes.com article). That’s a ten-fold increase in our daily ad load in my lifetime, and it’s mostly the internet’s fault. For the first decade or so, advertisers were slow to shift from traditional media to the internet. But as consumers’ attention shifted online, the ad dollars had to follow. Today, there are 5.3 trillion display ads online each year, and the ad-based revenue model has spawned some of the most valuable companies in the world. Two of the ten largest companies by market cap—Alphabet and Meta—derive the vast majority of their revenue (approximately 80% and 97%, respectively) from advertising. Three more—Amazon. Apple, and Microsoft—are gatekeepers eager to exploit their positions of power to either handicap their ad-dependent challengers and/or expand their own advertising revenue.

What makes the modern advertising-industrial complex so pernicious is not only its scale but its ubiquity and invasiveness. Like any business, Meta, Google, and any other ad-supported business must grow. Core to their immense valuations is the high rate of that revenue growth. And there are only three growth levers for ad-dependent businesses: more users, more ads per user, or more revenue per ad. As these platforms have plateaued in terms of user growth, they are forced to resort to the other two levers, with more ads and better targeting. The result is a a disconcerting harvesting, aggregation, and auctioning of our data at a truly epic scale.

The ad-based business model also seems to be breaking down for advertisers. Struggling to stand out in the tsunami of pitches, advertisers are complaining about decreasing ad effectiveness, plummeting click-through rates, and poor return on ad spend (ROAS). The entire advertising industry feels like it is simultaneously facing diminishing returns and a tipping point.

Against this backdrop enters Artificial Intelligence.

Though the novelty of ChatGPT has captured our collective imagination, the real prize of AI is that it will usher in a fundamental sea change in how the internet works. Today, we mostly “search” or “browse” for content online. Both of these routines offer convenient opportunities for advertisers to interject themselves—Google’s search advertising and Meta’s news feed advertising being the dominant examples of each, together accounting for over half the entire online ad market. AI threatens to disrupt these predominant models with a new way of finding information: curation. Rather than searching on Google or doomscrolling our feeds on Facebook or Instagram, AI could just tell us what to read, watch, buy, and believe. What could possibly go wrong?

To a large extent, we already live in this world. Predictive analytics and fine-tuned algorithms dictate much of what we consume online, engineered to keep us optimally engaged and clicking. But AI will take this to the next level, with the potential to alleviate the burden of decision-making all together. Part of me wants this future now. Rather than going to Yelp or Angi and searching for a plumber, as I recently did, only to get inundated with texts, emails, and phone calls, couldn’t AI just do this all for me? Find the best vendors, requests quotes, negotiate terms, and schedule the work. I’d pay for that.

Yet another part of me dreads this future. I wrote an article recently in Writer’s Digest about the impact of AI on the publishing industry. My conclusion was that, though the concerns authors have about AI—including training engines on copyrighted material, failing to compensate authors, and inundating the market with AI-generated content—are real, the biggest threat to authors may be that their work will no longer be discoverable. This “discovery bias” in which only the books that are the biggest bestsellers with the largest promotional budgets ever get discovered while all others are marginalized into irrelevance is deeply concerning not only in publishing but every category of products and services.

Basically, this feels like it could break one of two ways (with heavy odds on the latter). If AI is to become an automated mechanism that we trust to curate and deliver the information we’re looking for without the distraction of advertising, it needs to be beyond reproach. We need to trust its unbiased recommendations completely. If on the much-more-likely other hand, AI becomes yet another tool exploited by deep-pocketed advertisers to ensure mindless selection of their product or service, we could be headed for a world unburdened from the Paradox of Choice only to become more like Minority Report.

Ads suck, but as manipulative, deceptive, and invasive as they may be, at least we still possess free will. Many of us are so overwhelmed by choice, we don’t even realize when we gladly give it up—succumb to the algorithms serving us the next dopamine hit. But if we allow AI to narrow our decision-making aperture, we may not even know other options exist.