AI Advertising Doesn’t Have to Suck

Since OpenAI’s announcement that advertising is coming to ChatGPT, there has been no shortage of buzz on this topic within the ad community. While most people agree the initial ad examples and assurances about data privacy seem straightforward, innocuous and familiar (harking back to early Google search advertising), nearly every conversation ends in skepticism and a general sense that the “enshittification” of the LLM experience has begun. This sentiment is most vividly and publicly expressed by Anthropic’s four new Superbowl ads entitled Deception, Violation, Betrayal, and Treachery. Ouch.

Hard truth: this is the culmination of decades in which nearly every new product supported by advertising eventually went down the path of clutter, annoyance, creepiness, and diminishing trust. It’s also probably why OpenAI CEO Sam Altman just 18 months ago termed advertising a “last resort.”

And yet, despite widespread complaints about advertising, the economic reality remains that most consumers tolerate ads and prefer paying with their time and attention rather than their wallets. Consumers expect digital services to be “free” or subsidized by advertising. It’s why nearly half U.S. streaming subscriptions are ad supported and FAST has been on a hot streak for years.

But tolerance isn’t satisfaction. While the model of pushing more ads, capturing more data, and targeting more finely has created immense value for certain players and parts of the ecosystem, too often it’s been at the expense of consumers’ experience and trust, driving those who can easily afford it, the very consumers advertisers most want, to pay for ad-free.

It Doesn’t Have to Be Like This

We would be myopic to assume the only way to build a successful LLM ad business is to do it the Facebook Ads way or the Google Search way. Are those wildly successful businesses? Yes, but that doesn’t mean we stop innovating or seeking a better alternative. The fastest way to get somewhere used to be a horse, until Karl Benz finally threw away the model and started with a blank sheet of paper.

Can the ad business break the pattern that seems to define it? I sure hope so, and the start of ads on a game-changing technology that touches just about everything, feels like a momentous opportunity.

The way forward starts with a simple brief: Make it better than what’s been done before  Maximize Customer Lifetime Value. Prioritize consumer control. Collect only data you need to make it work versus everything you can possibly get.

I'm willing to bet the LLM platform that chooses this path gains real differentiation and another revenue stream. Higher consumer trust translates into greater reach, engagement, and quality consented signals, all of which drives ad efficacy and revenue. A thoughtful experience drives stickiness. This is a bet that could pencil out meaningfully for the company brave enough to make it.

Playing the Long Game

None of the leading AI companies with their multibillion-dollar investments are building a short-term business, so why let short-term thinking dominate the ad strategy? Anthropic’s Claude has a thoughtful 57-page document about its soul. Microsoft publicly released a detailed Responsible AI Standard, inviting outside feedback to build better industry norms. These companies are building platforms with serious intentionality, why not the ad experience?

Perhaps part of the reason is because many companies get into advertising in a hurry in response to financial pressure. Anthropic CEO Dario Amodei’s comments remind me of Netflix’s path into advertising; Emphatic and extended resistance followed by a sudden reverse. If the plan is to eternally sail the good ship “Never Ads” through the calm waters of subscription/enterprise revenue, my counsel would be to plan now for financial headwinds and establish an internal skunk works to figure out a better way to do ads. Thus, when financial pressure ratchets up  (which may be inevitable) they aren’t scrambling.

Regardless of whether these companies move into ads now or later, the thing is to be thoughtful and deliberate. For once, I want to see a company play the long game around advertising and unwind the short-sighted thinking where platforms benefit at the expense of consumers. I’m not talking about altruism. Companies exist to generate profit, but strong profits and poor ad experience need not go hand in hand.

What It Could Look Like

So, what to do? Let’s lower the cynicism and do what an industry that prides itself on creativity is supposed to: Create something new. This is a moment for innovation and perhaps even a new digital ad paradigm, one that puts consumers in control, gives them transparency, and earns their trust.

Imagine if, as you set up your LLM account, it flat out talked to you in lay terms about the value exchange inherent in its ad model and gave you some choice in your ad profile and ad experience. You could permission or actively choose the categories of ads you are most interested in. It could tell you what anonymized data (age, gender, IP address) was or was not being used for advertising.

It could offer you the choice to answer some specific questions to be used for ad targeting or measurement in exchange for some value like faster load times, more queries, access to deeper thinking, or lower monthly fees.

Certainly, it would be trust-building if the introduction of ads gave users more control over how their data is used, what ads they see, and the ability to change those settings later. Not to mention, the option to establish ground rules for the data they already shared prior to the launch of advertising.

Consumers could initiate a conversation directly about a potential purchase and an agent could ask, “Do you want me to show you deals from some of my advertisers?” Show a deal a consumer asks for that meets their needs and AI becomes a genuinely helpful assistant. You might be surprised how much consumers value clear choices, delineation between ads and content, and being given a degree of respect, control and empowerment.

A few years ago I ran a company called true[X] that pioneered opt-in engagement ads. Viewers were given the choice to engage with a brand and give better attention in exchange for specific value like removing an entire ad pod or getting preferred access. The value exchange was transparent and directly tied to the experience the user intended to access. On average, these opt-in ads delivered 5x more brand lift than standard full-episode video ads.

Data Matters. Just Not All Of It.

Before we assume LLM platforms will need to harvest reams of personal data to drive ad revenue, it's worth noting that in premium streaming CTV, the most common programmatic signals remain relatively coarse: IP addresses for household-level targeting and geography, inventory authentication signals and device identifiers. In other words, ad effectiveness does not necessarily require deep behavioral surveillance.

Furthermore, driving ad efficacy and ad revenue does not always require precise targeting and personalization. Some of the most effective (and expensive) advertising is about delivering the same message to massive audiences simultaneously. Empirical research from Byron Sharp demonstrates that over-emphasizing targeting can diminish a brand's mental availability, and ROAS over time. Similarly  John Dawes showed that 95% of B2B buyers are out-of-market at any given time, meaning narrow targeting misses the vast majority of future buyers.

The Upside 

These are just a few ideas. The tip of the iceberg. Advertising is meant to be a creative business, if more of us put our heads together there is so much more. AI capabilities are advancing at a wild pace with new possibilities rapidly emerging. The key is starting from a clean slate and with a consumer-first perspective, not retrofitting old advertising models onto new technology.

The upside to getting this right is real.

Consumers get something they want, transparency and control over their ad experience and their data.

Marketers and brands get the halo of supporting a platform that gives viewers control as well as access to permissioned, clear, user-initiated targeting and measurement signals. A combination of the high engagement these LLMs create, signal accuracy, and consumer goodwill would drive ad effectiveness. Not only that but it might even save on some of the ad tech tax of a third-party data broker.

For LLM’s and soon to be hatched agentic ad businesses, there is opportunity to drive business results, and yes profits, from transparency, value exchange, calibrated use of data, along with innovation in ad measurement, creative, and ad format. The challenge is to avoid short term thinking and the easy button of what’s been done before. Outsize returns require differentiated effort.

The LLM company that builds its ad business with transparency, control, and genuine value exchange could prove something remarkable: that advertising can make things better and drive the bottom line.

Posted at MediaVillage through the Thought Leadership self-publishing platform.

Click the social buttons to share this story with colleagues and friends.
The opinions expressed here are the author's views and do not necessarily represent the views of MediaVillage.org/MyersBizNet.