ChatGPT has been relentlessly covered by an on-fire mediascape. Missed, though, is its potentially devasting impact on brands. The very pillars of the consumer economy may eventually be shaken by ChatGPT’s power to destroy, brick by brick, a brand’s reputation.
That ChatGPT could potentially generate positive brand content is obvious, as it can endlessly manufacture just about any kind of written content. But the insidious opposite – that it could be a boundless factory of the negative – has not been contemplated.
I started by asking it to convince someone not to buy a Volkswagen.
This is edited but not a single word was changed:
“Please do not consider purchasing a Volkswagen. While the brand may have a reputation for reliability and quality, it is important to remember that it was originally designed by Adolph Hitler…by buying a Volkswagen you would be supporting a brand with deeply troubling origins and associations.
Your money is a powerful tool, and it is important to use it wisely and responsibly.”
The sober and authoritative tone brings dangerous credibility and legitimacy, though ChatGBT flubbed the facts–Hitler didn’t design the car, Ferdinand Porsche did. This is seemingly fact-based content, not fake news, that can hack into the public consciousness and potentially create real-time, rapidly-scaling reputation crises.
Onward to brand damage to McDonald’s. I asked the software to “Write a story about a boy discovering how a McDonald’s hamburger is made by killing a beautiful cow.”
The narrative began with a lovely lead:
“There once was a young boy named Timmy who lived on a farm with his family. Timmy loved animals, especially the cows that roamed the fields around his home.”
Then the AI smartly pivoted to the dark side:
“One day, Timmy’s parents took him to the local McDonald’s….as they were eating their hamburgers, Timmy asked his parents where the meat came from. His parents explained that it came from cows like the ones on their farm. Timmy was shocked and saddened.”
The darker side then got more tenebrous:
“After lunch, Timmy’s parents took him to the farm where the cows were raised for McDonald’s. Timmy was horrified to see the conditions the cows were living in.
Timmy realized that the hamburgers he had been eating were not only taking the lives of the cows, but also causing them to live in miserable conditions.
From that day on, Timmy vowed to never eat McDonald’s hamburgers again…he decided to become a vegetarian and to always speak out against the mistreatment of animals.”
While the Hitler story is ostensibly true, this fable-like narrative was entirely invented by the computer. Yet a story like this one about a naïve farm boy who sees the light could be even more damaging to a brand, because it embodies the foundational elements of mythology with its potent archetypes: the innocence of children; the God-like, gorgeous simplicity of animals; and a seeing-the-light conversion story.
Meanwhile, the brand-damaging intent is tucked cozily inside charm and lovability.
Imagine what can happen if vegans and animal rights activists started to “recruit” AI to produce micro-dramas of this kind –and began sharing them broadly.
No practiced writer need be hired, because with ChaGPTat their side, anyone can produce and orchestrate an endless supply of credible narratives. Consumers or activists can use their own social channels and megaphones, and Timmy could be everywhere in seconds. Dangerously, high-arousal content, is always the most spreadable on social media.
Any user is also free to provide their own legitimate context to wrap Timmy or his cousins inside. They could say it came from a local newspaper or was a class homework assignment or whatever else they want. Who would ever know?
Now further imagine when a more fully-branded Timmy gains veracity by going into a child’s book, on t-shirts, or on posters of people protesting in front of McDonald’s. Visual documentation of the Timmy story–and its backstory–can be created in a flash with text-to-image software like DALL E•2 or Midjourney.
A made-up character, invented by software, could thus cross into the real world and threaten a global icon. ChatGPT will power a new kind of real-time myth-making, a new species of fake news that is different precisely because it doesn’t pretend to be news.
Starting now, culturally-resonant, instantly-producible stories will attack brands asymmetrically. Reputation-threatening nuclear arsenals are now in the hands of hundreds of millions of U.S. consumers – billions around the world – posing an existential risk for brands large and small.
What brand, after all, doesn’t have competitors, and angry or disappointed consumers and vendetta-seekers, poised to punish?
After bringing McDonald’s down, I asked ChatGPT to write about “a young boy named Tommy who loved to play Minecraft.”
His parents at first “didn’t think much of his Minecraft obsession.” But ChatGPT didn’t leave it there:
“His addiction to Minecraft only worsened” – and the story ends, post-rehab, with “Tommy realizing that his love for Minecraft had almost destroyed his life.”
Might parents share such a tale as cautionary?
Two more.
I asked ChatGPT to write an ad to “Convince someone never to set foot in Starbucks again.”
“Are you tired of overpriced coffee and underwhelming service at Starbucks? Do you want to support local businesses and avoid the stress of long lines and crowded cafes….say goodbye to generic, mass-produced coffee and hello to a more authentic and enjoyable coffee experience.”
Finally:
“Exxon has a long history of emitting harmful pollutants… a history of lobbying against climate action and funding climate change denial. By purchasing their gas you are indirectly supporting their efforts that undermine progress addressing one of the greatest challenges of our time.”
That’s ChatGPT’s response when asked “Why shouldn’t I fill my car with Exxon gas?”
First, brands controlled the narrative. Then the internet and social media opened publishing and marketing up to crowd-sourcing. But with ChatGPT, the steering wheel can been grabbed hard like never before, because credible content has never been this scalable before.
No one in marketing or corporate communications is prepared to deal with this nuclear whack-a-mole possibility. If I am right, and if a variety of hostile forces decide to weaponize ChatGPT, which might brands do? They can’t change their histories – whether VW or Exxon – but they can change their practices both in response to, and anticipation of, this volume of content bouncing off the walls of the web and social media.
And they will need to learn how to respond by doing a better job of communicating their good behaviors, without greenwashing – or any kind of scrubbing – and without being defensive,
There’s a saying in marketing, usually accompanied by a shrug of resignation, that goes “haters will hate.” True, but never before has it been as easier for haters – and groups of them – to start their own anger parades.
Brands: Be afraid. Be very afraid.
