Using AI signals that you don't care... even when you do

The human brain is a remarkable thing. All that filling in the gaps¹, making sense of the world… A marvel of natural engineering. Its ability to pick up patterns easily is extraordinary - even if those patterns were not planned in the first place. The sky gave us stars, we saw animals; Hermann Rorschach gave us random inkblots, we saw objects; heck - old telescopes gave us grainy patterns on Mars, we saw canals.

The most recent pattern all of us are painfully familiar with is, of course, AI writing. Telltale signs are ubiquitous, obvious and loathed enough to the point that mistakes and deliberate imperfections are now in vogue.

Yes, there’s a tool for that.

It’s just fascinating to observe how humanity is reacting to AI. I’m interested in perception - that oh-so-subjective matter that the logic-driven homo economicus just can’t fathom. In a textbook world, perfect output - which is theoretically what AI writing is - should please everyone. Why doesn’t it?

---

I feel there are three reasons we’re put off by AI writing. When we see AI writing, we perceive:

• A lack of effort (“creating this was too easy for you”)

• Lack of original thought (“you didn’t have any ideas of your own, so you turned to ChatGPT)

• A lack of faith in one’s own skill (“and if you don’t trust your own abilities, why should I?”)

The thing is, all of the above could be wrong. Writing a quality prompt takes effort. Someone might have fed a hyper-original thought into AI just for polish. But the thing is, seeing those telltale AI signs in the output negates all that. It might be the most original thought in the world, but the minute one sees that “it’s not this it’s that” trope, we disregard it.

All three points ladder up to the same thing: A perceived lack of value of the recipient’s time and attention - the most important currencies in today’s time.

In effect, when you encounter AI writing, what you feel is: Lazy stuff that lacks substance, by someone who think they themselves fall short, actively disrespecting my time. OK, you may not react as strongly, and not all AI writing is bad. But what’s that old joke about 98% of lawyers giving the rest a bad name?

---

No, brands… Pls don’t.

It’s bad enough when individual people do this - it makes LinkedIn and some Substacks intolerable - but it’s unforgiveable when brands do it.

If I see AI-generated images of food at a restaurant, I just don’t order. This is not me taking a moral stance - it’s just a lack of certainty. Is this place not confident enough of their own food to put pics out there? I mean, it’s their own food - not, say, conceptual architectural designs that don’t exist in real life yet. Do they not care enough to hire a photographer or just even put imperfect photos out there themselves? Suddenly, all those restaurants with slightly-unappealing-but-very-real photos seem a safer bet than those using generated images.

I felt this most acutely with Urban Company - it felt a little strange to see AI-generated images for intimate, high-touch services.

Recently, Colgate was skewered for its use of AI imagery where even the product label was not clear:

I am sympathetic to the cost factor - not everyone has the means to shoot or buy stock images. And I certainly do not begrudge my neighourhood kebab cart for using a generated logo. But these are consumer brands with deep pockets / investments touting premiumness and trust (Urban Company is as 1% a product that exists).

I am shocked that Colgate did this with their own product - are you telling me there existed no brand photoshoot or assets?

Sagacious marketing observor Kartik Srinivasan has often lamented brands using AI-generated images - I don’t think the problem is using AI as much as the lack of attention - didn’t someone bother to check the full output?

To me, that is egregious. What are all those 300-page brand books and 5 levels of approval for?

Those things I said above about perception? They apply to brands too.

When I see a huge brand using an obviously AI-generated image, it doesn’t signal a lack of budget. It signals “I don’t care enough about you to use an actual image or even fix this generated one”. I’d argue the impact is even stronger - you may forgive your colleague for using AI to polish up his post, but a multi-billion-dollar company? That just sends the wrong signal. These two comments on the Colgate post sum up my feelings nicely:

Let’s not even get started on how you would react to AI-generated customer care mails or useless chatbots on a company homepage. Your angry email met with a GPT response will exacerbate all those stressors, and then some.

I get it, it boils down to savings. Everyone - even the huge companies - are under pressure to cut costs. The question is whether those savings are substantial enough to offset even a momentary drop in perception, let alone introduce a new barrier to purchase.

---

Of course, AI is useful

Before the AI doomer camp gleefully accepts me as their tribe, let me be the first one to say no thanks. I think there is a lot of good use of the technology, and I think there are legitimate use cases.

It’s a remarkable research tool. Companies have developed new flavours, blends and product lines. Not to mention everything happening in drug discovery and protein synthesis. And certainly, there are creators who’ve used AI in very interesting and engaging styles.

The difference though is - that shows effort. There is a layer of human curation and validation atop them.

And in my view, that’s what separates the think line between purely AI music, and music that uses a lot of technology (just look at the amount of technology - from DAWs to guitar effects pedals to amplification - that goes behind an average rock song. Surely someone from a past generation would say using all that technology is as bad as some purists today would say using Suno is).

I’m fond of abstracting out technology to get to what humans are actually doing to each other. Here is, abstracted of AI, what using AI signals in various situations:

• A professor looking at an AI-worded assignment: This student cheated

• Seeing an AI-generated LinkedIn post: Gosh, more rubbish from these corporate types desperate for attention

• Seeing an AI-generated product video: Is this thing even real? Does it exist? Can I trust this company?

• Perusing a high-value service and seeing AI content on the site: Umm… If they’re too lazy to write on their own, will the service be good, especially for what I am paying?

AI itself isn’t the culprit - human arrogance or hubris is… as it always has been.

---

One of my favourite books is Rory Sutherland’s Alchemy - a key takeaway for me being that humans value effort (to paraphrase - you wouldn’t propose to someone over text).

We’re really evaluating the effort, not necessarily the output.

I mean, this song sounds nice on its own:

But then when you see how it was actually done, your respect for it goes up.

Really, we’re all just evaluating effort. So when a professor gets the ‘perfect’ essay, she starts questioning whether AI was used and possibly gives a bad score to the student. It’s the reason why we still tune into human sports, despite robots being able to play a technically better game. That imperfection, that humanity, still appeals to us.

Now I am not going to encourage deliberate typos, or shunning AI tools altogether - but what I do encourage is quirk, and building your own style, your own voice.

And you know who helps you do that? The 6% Club.

That’s right, this entire essay has just been content marketing to peddle my course, which is into its 9th cohort, and has mentored over 300+ participants. This entire ooh-AI-writing-is-silly diatribe has just been so that you can come and find your own category of one with 4 live sessions, hands-on exercises and masterclasses by Krish Ashok, Amit Varma, Ria Chopra and Mahima Vashisht. Yes, this full post taking a sh*t on poor multibillion dollar brands is just so that you can take advantage of the 10% discount we offer all readers of this newsletter, which you can avail when you ping Utsav on WhatsApp to join the program.

Let’s see AI come up with an idea like that.

Wow what a deux ex machina ending that was.

Stay human, I guess. It’s good for business!

Chuck

1. Did you know there’s a literal blind spot you have now? There’s a part you are seeing which you’re actually not ‘seeing’ - your eyes can’t cover it - your brain is filling up the parts. True story!