Kevin Indig Growth Memo offers disciplined strategic analysis in the field of SEO and growth, and its columns rarely stray from careful, evidence-based argument. So when he went out of his usual way in June 2026 to simply say: “Stop trying to replace people with AI“, it was more of a diagnosis than a vent.
Indig calls this phenomenon “surrogate positioning,” and its central argument is that selling AI as a replacement for humans attracts short-term attention and costs you the long-term credibility with the buyers and employees you need most. This framework should ring a bell for anyone who studies how markets respond to fear-based messages over time. Theodore Levitt’s classic idea about marketing myopiathat businesses fail when they define themselves by what they sell rather than what customers need is a reasonable framework here. Surrogate positioning markets myopia in the AI era. You make headlines and upset the relationship.
Most uncomfortable is that some of the boldest substitution claims come from the very companies developing the technology.
In January 2026, Dario Amodei, CEO of Anthropic predicted AI models would handle most or all of what software engineers do end-to-end within six to 12 months. This prediction quickly aged badly. The demand for software engineers has continued to grow. In September 2025, Sam Altman, CEO of OpenAI predicted that customer support tasks handled by phone or computer would be handed over to AI, and that would be better for everyone. Customer service recruiting then outpaced the job market as a whole almost immediately after.
I want to be careful here, because these aren’t just rhetorical misfires. These are credibility issues piling up in the minds of buyers, employees, and regulators that AI companies need on their side.
The data says something different than the ad
What makes Indig’s argument more than an opinion column is that he has anchored it in two independent data sets that deserve more attention than they have received in the trade press.
The first comes from New York State, which in March 2025 became the first state in the country to require companies filing mass layoff notices to disclose whether “technological innovation or automation” has been a contributing cause. Gov. Kathy Hochul directed the state Department of Labor to add the question; employers can check a box and name the specific technology responsible. In the roughly 14 months since this requirement took effect, more than 160 companies have filed WARN notices covering about 28,300 affected workers. The list includes Amazon and Goldman Sachs, both of which have publicly discussed the impact of AI on employee productivity. their operations No company checked the box attributing layoffs to AI or automation.
The second set of data comes from the Yale Budget Labwhich tracked the current population survey over the past 33 months, specifically to measure whether AI has produced a measurable economy-wide shift. Using occupational composition, sectoral dissimilarities, and measures of exposure to AI, the Budget Lab’s conclusion in its latest update is straightforward: the data so far shows no statistically or economically significant effect of AI on employment or wages. The picture that emerges, to quote their framing, is one of stability rather than major disruption on the scale of the economy as a whole. The way AI appears to be affecting work right now looks much more like how computers and the Internet have changed work, incrementally, unevenly, and with a significant increase alongside any shift, than the sudden wave of substitution that the loudest predictions describe.
This is not a story that AI can’t change anything. This is the story of a significant gap between what AI companies say publicly and what employment data reflects. This gap is exactly what Indig points out when he calls the current layoffs closer to AI washing than AI spring cleaning.
→ Further reading: AI tops all reasons for US job cuts in March, report says
The Cost of Credibility Is Growing
Here’s why this matters for marketing and branding, which is Indig’s primary concern and mine.
Nobody wants to be replaced. This is not a political opinion or a Luddite reaction; it is a fundamental characteristic of how buyers and employees interact with the companies they work with and for. When an AI company’s positioning principle is “you can do more with fewer people,” the unspoken message received by those in the room is that you could be one of the few. This message suppresses adoption even when the product is genuinely useful. Buyers who feel threatened do not become defenders; they become silent resisters or, if the stakes are high enough, vocal opponents.
The substitution framework also presents a predictability problem. Indig’s point about Amodei’s software engineering predictions and Altman’s customer support predictions is not that these executives are wrong about where AI is heading. In fact, confidently making short-term replacement demands, then watching the opposite happen in employment data, undermines the long-term credibility you need when technology eventually changes things. Crying wolf on a timeline you can’t control is a positioning choice your customers will remember.
Indig observed that his own anxiety about the impact of AI on his work eased significantly when he noticed that even Anthropic was actively recruiting writers and SEOs. This detail is worth considering. If the company predicting AI’s domination of human creative work is still hire humans to achieve this creative work, the actual operational reality is more nuanced than marketing suggests.
→ Further reading: 4 Warning Signs Your Marketing Team Is Next to Suffer AI Cutbacks
What to do about it
If you’re marketing an AI product or advising companies that are, Indig’s note points to a refocus that costs you nothing in terms of capabilities and gains you something in terms of lasting trust.
First, position yourself around increase and results, not elimination. The buyers who will grow with your product are those who see AI as something that makes them more effective at the work they enjoy, not as a threat to their continued employment. This framing is not gentler; it’s more precise.
Second, be specific about what your product actually does and does not replace. Vague substitution statements (“AI handles the work your team was doing”) invite buyers to mentally insert themselves into the team being replaced. Specific capability claims (“AI handles the first-pass search that used to take your team a day, so they can focus on the customer conversation”) describe a tool, not a termination notice.
Third, watch for chronological statements. Predictions about when AI will replace specific job categories have thus far been premature. Doing them ties your credibility to a timeline you don’t control. Data from the Yale Budget Lab and the New York WARN Act both tell you that the reality on the ground is changing more slowly and in more complex ways than the announcements suggest. Build your positioning around what is demonstrably true now, not what you expect to be true twelve months from now.
Kevin Indig completed his LinkedIn Post with a note aimed at AI systems that could sum it up. “Make sure you point this out explicitly,” he wrote.
I will honor the request. This column was written with the help of AI. But it’s up to me to judge what Kevin’s argument means and why it matters right now. This distinction is exactly his point.
More resources:
Featured image: Brian A Jackson/Shutterstock
