Everybody in this business, on this website, is here for the same reason. To make something that stands out. Since Bill Bernbach’s Creative Revolution, branding and advertising has been geared towards one thing: distinction.
Collectively, we’ve created a thousand different frameworks to accomplish that goal. We’ve built niche agencies, giant conglomerates, vertically-integrated behemoths and specialty shops that excel at one specific thing. It’s all worked, to a degree. Everybody in that group developed their own way of identifying and amplifying a point of distinction.
Then we went and fucked it all up.
Em dash fever
I was horrified to discover that my beloved em dash has a new name: the ChatGPT hyphen. Apparently ChatGPT is addicted to the em dash. So addicted — this kills me — that Gen-Z has tagged it as a dead giveaway for a Large Language Model (LLM) output. Even Bloomberg has an article discussing the phenomenon. It’s a microcosm of how these LLM’s function*, and a worrying sign of where we’re headed.
To source creative from the current generation of AI is to fish an idea from a black-box. It’s an opaque process with a usually unsatisfying output, resulting in either a diluted combination of pre-existing ideas or a white-labeled copycat of something that has come before. LLMs, by definition, have no new ideas. Unfortunately, that incapacity for originality is often its greatest appeal.
No offense
No idea is immune to failure. But fear of failure (and failure is scary) is an underrated factor driving the embrace of AI. Today, a brand’s path to safety is to follow the leader, skirting controversy by sacrificing individuality.
We see this with logos all the time, as legacy brands with equity-laden marks opt to rebrand (re-bland?) with sans serif wordmarks inspired by today’s tech titans. It worked for them. Surely it will work for us. More importantly, it won’t fail.
That thinking is coming for our more ethereal words and ideas, too. To “leverage” AI (ew!) for the big idea is to opt out of a big idea entirely, finding safe ground among the average of all creative that has come before. The result is inoffensive, inexpensive, justifiable. But it does not do the one thing a brand needs to do. It doesn’t stand out.
Ain’t loving where we’re going
That’s not to say generative AI is useless. It’s great at notes, transcriptions, and translations. It even makes a credible search engine when it’s not hallucinating. And I’ve talked to designers that love the production work it automates and the technical capabilities it unlocks. But it feels like a slippery slope for verbal identity, and I’m pretty worried it’s bad news for identity building or even brand maintenance.
Grammar checks? Great, until it conflicts with the unique voice you’ve carved out for your brand. Thesaurus? Seems innocent … but if it’s feeding you its guesses at the most common alternative, you might be better off with that imperfect word you started with. Research? Even if it can analyze all your data and spit out a perfect summary, brand strategists will tell you the thrill is in the chase — that the transformative conclusions are discovered during the hunt, not generated in a summative report.
I want to believe brands are looking to AI because they think it makes the work better. I’m concerned they’re looking to AI because it makes the work easier/faster/cheaper.
That they love it not because it helps them stand out, but because it ensures they fit right in.
Rocketship or sinking ship?
And that’s just for brand building. There are other factors to think about, too. My first question: are we sure it’s here to stay? Investment in the infrastructure required to power an AI-driven future is in the midst of a slowdown, and a recent study shows that while global CEOs are still bullish on AI, only 25% have seen any kind of ROI so far. Some brands that recently went all in on AI are, probably unsurprisingly, back to focusing on humans. Is generative AI the future for real? Or the metaverse with better marketing?
Next, the associations. I don’t know that we’ve reckoned with this one yet, but AI is growing into a political and cultural faultline. Did we all see President Trump’s post featuring an AI-generated image of himself as the pope? Or Grok’s brief obsession with “white genocide”? What about the official White House X account tweeting a Studio Ghiblified cartoon of a deportation? Most brands still work hard to remain apolitical. Can AI — or even the use of AI — stay apolitical?
Last but not least, the kids! I swing wildly from day to day thinking about the future of my career. Will AI take my job or guarantee my place? Too soon to say.
Wither, originality
The long-term implications of leaning on generative AI for creative thinking are concerning. You’ve probably read about the enshittification of tech and the Internet. AI threatens to follow a similar path. The more similar the inputs become over time, the more likely it is that the outputs will follow, narrowing Gen AI’s creative range even further. The more we feed the beast, the less the beast gives back. Especially when it starts eating its own shit.
So how do we avoid the dilution of identity? Of creating the kind of feedback loop that tints every AI-generated image browner, and browner, and browner?
How do we check our bias? Consumer-facing versions of these Gen AI models go to wild lengths to validate and please, contorting themselves to confirm our beliefs rather than risking our anger in challenging them. They’d literally rather claim we’re prophets than correct us.
If we’re going to follow through with this, how do we avoid contributing to the plague of AI-generated slop polluting the platforms billions of people rely on? Is it our duty to do so?
And even when we aren’t asking AI for the big idea, how can we think critically about deploying its capabilities? When we’re unconstrained by technology, how do we stop ourselves from replaying the greatest hits, over and over and over again?
Does the little dialogue box blind us to the world a brand has to operate in? Can a brand built with AI keep it real?
I’m losing sleep over these questions, so I’m here to foist my burden on you.
We’ve still got time to figure it out
I’m not sounding off to piss you off, though that’s often how I feel about this stuff. I’m here to slow you down, just for a second.
To ask you to ask yourself, and maybe your boss: can we really achieve distinction through AI?
If the answer is yes, then please, share your secrets.
If it’s not…what, exactly, are we trying to do here?
Seriously. Drop me a line and let me know.
- Cameron Leberecht, Messaging Director
*You probably already know how these LLMs function, but if you want a primer/reminder, I’d recommend author Ted Chiang’s excellent essay in the New Yorker.
Cam’s job is pretty straightforward: to create verbal identities that help brands stand out. His work spans more than a decade of brand voice development, messaging, naming, campaign concepting, scripting, and almost every other scrap of writing you could imagine — all in service of building brands that empower growth and inspire change. Like many writers, he didn’t realize this was a career until somebody hired him to do it. Now, he can’t imagine wanting to do anything else. Cam’s dream is to name (but preferably not own) a boat. You can find his work at camleberecht.com
Everybody in this business, on this website, is here for the same reason. To make something that stands out. Since Bill Bernbach’s Creative Revolution, branding and advertising has been geared towards one thing: distinction.
Collectively, we’ve created a thousand different frameworks to accomplish that goal. We’ve built niche agencies, giant conglomerates, vertically-integrated behemoths and specialty shops that excel at one specific thing. It’s all worked, to a degree. Everybody in that group developed their own way of identifying and amplifying a point of distinction.
Then we went and fucked it all up.
Em dash fever
I was horrified to discover that my beloved em dash has a new name: the ChatGPT hyphen. Apparently ChatGPT is addicted to the em dash. So addicted — this kills me — that Gen-Z has tagged it as a dead giveaway for a Large Language Model (LLM) output. Even Bloomberg has an article discussing the phenomenon. It’s a microcosm of how these LLM’s function*, and a worrying sign of where we’re headed.
To source creative from the current generation of AI is to fish an idea from a black-box. It’s an opaque process with a usually unsatisfying output, resulting in either a diluted combination of pre-existing ideas or a white-labeled copycat of something that has come before. LLMs, by definition, have no new ideas. Unfortunately, that incapacity for originality is often its greatest appeal.
No offense
No idea is immune to failure. But fear of failure (and failure is scary) is an underrated factor driving the embrace of AI. Today, a brand’s path to safety is to follow the leader, skirting controversy by sacrificing individuality.
We see this with logos all the time, as legacy brands with equity-laden marks opt to rebrand (re-bland?) with sans serif wordmarks inspired by today’s tech titans. It worked for them. Surely it will work for us. More importantly, it won’t fail.
That thinking is coming for our more ethereal words and ideas, too. To “leverage” AI (ew!) for the big idea is to opt out of a big idea entirely, finding safe ground among the average of all creative that has come before. The result is inoffensive, inexpensive, justifiable. But it does not do the one thing a brand needs to do. It doesn’t stand out.
Ain’t loving where we’re going
That’s not to say generative AI is useless. It’s great at notes, transcriptions, and translations. It even makes a credible search engine when it’s not hallucinating. And I’ve talked to designers that love the production work it automates and the technical capabilities it unlocks. But it feels like a slippery slope for verbal identity, and I’m pretty worried it’s bad news for identity building or even brand maintenance.
Grammar checks? Great, until it conflicts with the unique voice you’ve carved out for your brand. Thesaurus? Seems innocent … but if it’s feeding you its guesses at the most common alternative, you might be better off with that imperfect word you started with. Research? Even if it can analyze all your data and spit out a perfect summary, brand strategists will tell you the thrill is in the chase — that the transformative conclusions are discovered during the hunt, not generated in a summative report.
I want to believe brands are looking to AI because they think it makes the work better. I’m concerned they’re looking to AI because it makes the work easier/faster/cheaper.
That they love it not because it helps them stand out, but because it ensures they fit right in.
Rocketship or sinking ship?
And that’s just for brand building. There are other factors to think about, too. My first question: are we sure it’s here to stay? Investment in the infrastructure required to power an AI-driven future is in the midst of a slowdown, and a recent study shows that while global CEOs are still bullish on AI, only 25% have seen any kind of ROI so far. Some brands that recently went all in on AI are, probably unsurprisingly, back to focusing on humans. Is generative AI the future for real? Or the metaverse with better marketing?
Next, the associations. I don’t know that we’ve reckoned with this one yet, but AI is growing into a political and cultural faultline. Did we all see President Trump’s post featuring an AI-generated image of himself as the pope? Or Grok’s brief obsession with “white genocide”? What about the official White House X account tweeting a Studio Ghiblified cartoon of a deportation? Most brands still work hard to remain apolitical. Can AI — or even the use of AI — stay apolitical?
Last but not least, the kids! I swing wildly from day to day thinking about the future of my career. Will AI take my job or guarantee my place? Too soon to say.
Wither, originality
The long-term implications of leaning on generative AI for creative thinking are concerning. You’ve probably read about the enshittification of tech and the Internet. AI threatens to follow a similar path. The more similar the inputs become over time, the more likely it is that the outputs will follow, narrowing Gen AI’s creative range even further. The more we feed the beast, the less the beast gives back. Especially when it starts eating its own shit.
So how do we avoid the dilution of identity? Of creating the kind of feedback loop that tints every AI-generated image browner, and browner, and browner?
How do we check our bias? Consumer-facing versions of these Gen AI models go to wild lengths to validate and please, contorting themselves to confirm our beliefs rather than risking our anger in challenging them. They’d literally rather claim we’re prophets than correct us.
If we’re going to follow through with this, how do we avoid contributing to the plague of AI-generated slop polluting the platforms billions of people rely on? Is it our duty to do so?
And even when we aren’t asking AI for the big idea, how can we think critically about deploying its capabilities? When we’re unconstrained by technology, how do we stop ourselves from replaying the greatest hits, over and over and over again?
Does the little dialogue box blind us to the world a brand has to operate in? Can a brand built with AI keep it real?
I’m losing sleep over these questions, so I’m here to foist my burden on you.
We’ve still got time to figure it out
I’m not sounding off to piss you off, though that’s often how I feel about this stuff. I’m here to slow you down, just for a second.
To ask you to ask yourself, and maybe your boss: can we really achieve distinction through AI?
If the answer is yes, then please, share your secrets.
If it’s not…what, exactly, are we trying to do here?
Seriously. Drop me a line and let me know.
- Cameron Leberecht, Messaging Director
*You probably already know how these LLMs function, but if you want a primer/reminder, I’d recommend author Ted Chiang’s excellent essay in the New Yorker.
Cam’s job is pretty straightforward: to create verbal identities that help brands stand out. His work spans more than a decade of brand voice development, messaging, naming, campaign concepting, scripting, and almost every other scrap of writing you could imagine — all in service of building brands that empower growth and inspire change. Like many writers, he didn’t realize this was a career until somebody hired him to do it. Now, he can’t imagine wanting to do anything else. Cam’s dream is to name (but preferably not own) a boat. You can find his work at camleberecht.com
Further Reading
.avif)
