Why Consumers Love AI But Hate Paying for It

Recent research by eMarketer shows that around 50% of American consumers are unwilling to pay for AI features on their smartphones. This holds up to some netnographic research I've conducted for some tech companies in recent months. It's what I call the "Google Ritual Effect". So what is that and why is it key to understanding this reluctance to pay for AI features?

It's not really about price sensitivity, it's about an entire civilisation being trained for over to decades to expect asking questions and getting answers should be free. This is the friction for AI tools like ChatGPT, Claude, Gemini and other LLMs. And Google, along with other search engines like Bing and DuckDuckGo are feeding into this by providing LLM based summaries at the top of search results.

While consumers are starting to adapt to LLMs being part of their ritual search behaviours, to them, that is Artificial Intelligence (AI). The majority of consumers have little to no idea that AI is an umbrella term for a suite of tools and they generally don't care either. They love these AI tools, they just don't want to pay for them.

If we explore searching from a ritualistic perspective, we can begin to shed some light on what is happening. Searching is, as anthropologist Claude Lévis-Strauss would see it as a fundamental online cultural structure. A binary opposition between "knowing" and "not knowing" mediated through the ritual act of searching.

We might also apply some Marcel Mauss here too through his concept of gift economies. The seemingly "free" exchanges create powerful and deeply meaningful reciprocal obligations. The "gift" of free search crated a cultural expectation that information access should be unmonetised, even though the consumer became the product by reciprocating through having their personal information monetised by search engines. Over time, this resulted in a "digital potlatch" where search engines status was enhanced by giving away valuable services.

The consumer ritual of around search has become deeply embedded in consumers minds through ritual and behaviour and an accepted system of reciprocity. This, to the point where we can see in LLMs what Anthony Giddens would call a "diembedding mechanism" in that they remove information seeking from its familiar institutional context.

With LLMs and search engines, we see two autopoietic systems. The Google ecosystem is based on network effects and link-based authority. A sort of structural coupling with the web's information architecture. LLMs represent a fundamentally different systemic organisation, being more like what a biologist calls holobiotic systems, where the AI tool contains multitudes of information sources in a synthesised form.

Then there's the issue of friction and viscosity. The Google ritual is often much about "satisficing" where people develop heuristics for quickly identifying credible sources and relevant information. The result is low cognitive friction once learned and many millions have learned this. On the other hand, LLMs, despite being conversational in their interactions, introduce new forms of cultural viscosity:

• Authority anxiety: Without visible sources, users experience what I call "epistemic vertigo" (where one has a sort of detached perspective with an LLM)

• Prompt literacy: The shift from keyword-based to conversational query formulation requires new cultural capital

• Output validation: People must develop new heuristics for assessing AI-generated responses (something most people are terrible at if they have little to no skills in critical thinking and mental models.)

Free search has become a part of our cultural consciousness. So it's not just an expectation but an archetypal pattern. Payment for information access triggers what Rapaille might identify as survival-level cultural codes around scarcity and tribal resource distribution. I suspect David Graeber would call it a form of "baseline communism" with the assumption that information needs should be met through sharing rather than cultural exchange. Even though user data is the transactional price to be paid. Bit of a paradox there.

Each successful "free" search has reinforced the cultural expectation that valuable information access shouldn't require direct payment.

Yes, some people are more than willing to pay a monthly subscription fee for some AI tools. Some upwards of $200 a month. But these people are a different caste in culture. They are more likely to be better educated, have critical thinking skills and understand how to leverage mental models and frameworks. It is a fascinating cultural moment.