You Are Faking It
This week in AI & marketing
More than half (63%) of workers are lying about their AI skills. A recent study confirms this, though the researchers are nice enough to call it “exaggerating.” We are not that nice, I guess.
And you would think it’s the boomers and Gen X. The generations sitting in the armchairs and corner offices are the ones who lie the most, right? By a planetary distance, it’s Gen Z.
Now, we are not just less nice, but also very pragmatic. And, to be honest, old. We’ve seen a lot. So we assume innovation cycles are called cycles because they repeat. The one we keep coming back to here is the late-90s digitalization cycle. You know, when everything moved from paper to the cloud. There was a thing called the Internet, and everyone wanted to have something online.
What happened there? Jimmy, the 60-year-old boss, who’d seen a computer at a tech expo once but still ran his business through his Rolodex and his secretary, and who was an avid fan of Dale Carnegie self-help books, saw a blue ocean opportunity. He hired Richie, a 25-year-old his cousin recommended with these exact words: “Richie is very good with computers.” What could go wrong? Richie had no idea about business goals, and “very good with computers” meant he was great at copying files between two external drives and had once coded a snake game in Turbo Pascal. But he was amazing at talking his way into situations, so he quickly became VP of Tech at Jimmy’s company.
I’m sure we all know these stories by heart. I personally had two bosses like Jimmy, and I was Richie at least three times in my career. One of my Jimmys even quoted Dale Carnegie to me, multiple times, to my utter amusement.
But I was able to solve the problems that came up later. Even though I talked my way in (a nice way to say I lied), I was creative and driven, and I had to solve the issues. Otherwise, my fakery would have been exposed. In almost every case, I did solve them. When there is pressure, you can learn. And marketing is not brain surgery, so it is easy to learn. So we don’t think faking AI skills will be the problem. People will adapt, regardless of their skills. If they don’t, they will be cut out of the workflows (a nice way to say unemployable).
What we do think is that AI readiness reports and company skills and competency audits have almost no real value. Honestly, we never thought they had in the first place. Any audit is a waste of money unless it is connected to action and execution. Most of these “AI readiness audit” services were just moneybags for consulting and wasted expenses for management, letting them sleep on the problem.
Now we have a very solid reason why they mean nothing. The baseline data is fake. Most people lie about their AI skills. They either lie consciously, to keep or get their jobs, or they lie without even realizing it. The simplest example of the latter is the stupidest question on these audits: have you used AI before? They say yes, and the company reports, “95% of our staff have used AI, we have no problem.” And the usage means a free version of ChatGPT, where they chat about something of elementary complexity.
What’s the takeaway? You can see from this week’s news that AI will own the entire marketing cycle, because it will own the entire decision cycle for the customer. Which means advanced, real AI skills will matter. And many people lie about their skill levels. So the solution is to care zero and jump straight into action. Train the staff without a prerequisite audit, and not on generic AI skills, but on very narrow, niche-focused ones. Skills that matter to YOU and your business. Then let them build the workflows around it. It is all action, execution, going deep and vertical rather than shallow and horizontal.
And you have to do it now. This is hour zero.
— Torsten and Peter
Workers Are Quietly Faking Their AI Skills
Turns out the “AI-native generation” might not be quite as AI-native as LinkedIn would have you believe.
A new GCheck report found that 63% of workers admit to exaggerating their AI capabilities, with the number jumping to a staggering 80% among Gen Z respondents. Many workers say they confidently discuss AI in meetings despite lacking hands-on experience. Some even volunteer for AI-related projects they are not qualified to handle. In other words: the AI skills bubble has officially entered the workplace.
The interesting part is not the dishonesty itself. It’s the psychology underneath it. Employees increasingly feel trapped between two fears: looking obsolete if they don’t embrace AI, and becoming replaceable if they do.
This creates a strange workplace dynamic where everyone publicly champions AI while privately resenting or avoiding it.
One CEO quoted in the report called this “double distortion” — workers overstating their AI fluency while quietly undermining adoption internally.
As a result many organizations may be massively overestimating their actual AI readiness.
If employees are bluffing their way through AI conversations, then internal surveys, transformation initiatives, and even hiring pipelines may be giving management a dangerously distorted picture of capability. AI adoption is increasingly becoming a culture problem disguised as a tooling problem.
Google Wants Ads to Feel Like Part of the Conversation
Google’s AI Mode is slowly turning into something closer to a conversational operating system for commerce — and ads are becoming part of the dialogue itself.
At I/O, Google revealed new Gemini-powered ad formats for AI Mode, including conversational sponsored responses and “Highlighted Answers” embedded directly into AI-generated recommendations. Instead of banner-style placements sitting beside search results, the ad becomes part of the generated answer. Ask how to make your house smell like a luxury spa, and Gemini may recommend products inline while explaining why they fit your needs.
This is a very different mental model from traditional search advertising. Historically, Google sold clicks around intent. AI Mode aims to monetize the intent-resolution process itself. The recommendation engine, explanation layer, and shopping journey are collapsing into a single conversational experience. That has profound implications for marketers because persuasion increasingly shifts from keywords to contextual reasoning.
The risk, of course, is that brands lose visibility into why they are being recommended in the first place. As AI interfaces become gatekeepers for discovery, marketers may gain efficiency while simultaneously losing transparency. We are entering the era of “black box persuasion,” where optimization increasingly happens inside systems brands cannot fully inspect.
OpenAI Is Slowly Building a Real Ads Platform
Our months-long ChatGPT Ads saga continues.
Since the launch, advertisers complained about severe underdelivery, weak reporting, and inventory shortages so bad that some campaigns barely spent at all. But OpenAI has been shipping aggressively.
In recent weeks, the company added daily budgets, geo-targeting down to ZIP code level, expanded reporting, conversion tracking, CPC bidding, and dynamic CTA buttons like “Shop Now” and “Learn More.”
The direction is becoming clearer: ChatGPT ads are evolving away from “sponsored placements in a chatbot” into a proper intent-driven advertising ecosystem. And unlike social media, where users are often passive or distracted, ChatGPT conversations tend to happen during active decision-making moments. Several advertisers already report conversion efficiency approaching non-brand Google Search performance — which is remarkable considering how immature the platform still is.
What’s becoming increasingly obvious is that conversational advertising behaves differently from both search and social. The user is not browsing. They are thinking out loud. That creates a very unusual ad environment where relevance matters far more than interruption. The companies that learn how to participate naturally in those conversations — rather than simply inject messages into them — will likely dominate this new category.
Nike Is Treating AI Commerce Like a World Cup Moment
Nike is becoming one of the launch partners for Google’s new Universal Cart system inside Gemini and AI Mode, timed deliberately around the 2026 FIFA World Cup.
For the first time, a major global brand is treating “agentic commerce readiness” as a real commercial priority rather than an experimental side project. Consumers will be able to discover Nike products, assemble multi-item carts, and move toward checkout directly inside conversational AI experiences powered by Google.
The broader implication is that product catalogs are becoming marketing infrastructure. AI systems cannot recommend products effectively if feeds are incomplete, poorly structured, or missing machine-readable attributes. In the old internet, bad product data hurt SEO. In the AI-commerce era, it may make you effectively invisible.
Expect every retail executive to ask the same question over the next year: “What’s our AI commerce strategy?” And increasingly, the answer will have less to do with creative campaigns and more to do with structured data pipelines, inventory synchronization, and conversational discoverability.
Spotify and Universal Just Made AI Music “Legitimate”
Spotify and Universal Music Group announced a landmark agreement that allows fans to create AI-generated covers and remixes of participating artists’ songs — with licensing, attribution, and revenue sharing built in.
This is a much bigger story than “Spotify adds an AI feature.” The music industry appears to be shifting from fighting generative AI toward containing and monetizing it.
Instead of trying to stop fan-made AI music entirely, Spotify and UMG are building a licensed ecosystem around it. Their argument is simple: AI-generated music is coming either way, so better to create a controlled system based on consent and compensation than let unlicensed “AI slop” dominate the market.
That framing is likely to spread far beyond music. We are starting to see the emergence of a broader compromise between creators and AI platforms: not full resistance, but negotiated participation. The industries that survive this transition best may not be the ones that reject generative AI outright, but the ones that figure out how to wrap ownership, attribution, and economics around it before someone else does.
One uncomfortable question, though, remains unresolved: if AI-generated creative output becomes effectively infinite, what happens to human attention? In a world where one song can become “10,000 versions,” discovery will become the real scarcity.
Share with a colleague. Earn paid membership months.