Just in time for Christmas… well, mid-November… Coca-Cola launched a new take on its annual holiday ad campaign. This year’s entry has been entirely animated by generative AI technology. Actually, Coke hired three different AI studios – Secret Level, Silverside AI, and Wild Card – to create their own takes on a familiar Coca-Cola Christmas ad, and then widely shared their favorite installment far and wide.
If you’re curious, at least one of the alternates has also apparently popped up on social media.
The iconic soda maker also launched an original AI app – Create Real Magic – in which a virtual Santa asks you for personal data that he will use to make you a personalized magic snow globe. And you don’t even need to give him any milk or cookies!
Reactions from social media to the campaign have mostly stayed the same. The ad revisits a popular spot from the mid-1990s, called “Holidays are Coming.” Combined with the presence of a few signature polar bears, this seeming attempt to mix nostalgia for Coke Christmas ads gone by with an attempt to sell viewers on this new technology rubbed a lot of viewers the wrong way. This TODAY piece features quotes from across social media with comments like “terrifying,” “soulless,” “lifeless,” and “creepy,” with TikTok’s Dylan Pearce accusing the company of that most grievous of December crimes… ruining Christmas.
AI-generated parodies have also started to appear.
The company gave a typical corporate response, noting that it also plans to release non-AI Christmas ads this year – like this one, “Holiday Road” – and that they always play around with advertisement strategies on the “cutting edge” of marketing and innovation.
Still, it’s hard to not see some kind of sinister motive also at play here. The ad normalizes the use of AI in a major and irreversible way. A huge American brand has now released an AI ad. We’ve crossed that Rubicon. It’s no longer a theoretical “one-day” possibility. It’s a thing that happened and we all moved on. Should lots more companies quickly jump on board the bandwagon and release AI ads of their own, we could be in a new era where AI simply replaces conventional forms of marketing, whether the masses approve or not.
Two years after the first apps launched to the public and sparked our national conversation about the technology, Generative AI remains trapped in this same discourse loop. It’s also worth pointing out that this isn’t even Coca-Cola’s first experiment with AI advertising. In March of 2023, the company released “Masterpiece” in partnership with OpenAI, which used the tech to bring famous paintings to life.
Since March 2023 and even a bit before, AI proponents have been making a similar case. Like it or not, the technology is here, it’s improving all the time, and we’re right on the precipice of a new world in which computers do a lot of our work for us, sometimes at the cost of our own employment. After all, a major company now has an AI-generated ad campaign! Deny the inevitable future at your own risk!
Meanwhile, critics point out that the technology remains unreliable, inhumane, rooted in copyright theft, and designed to steal jobs away from creators and artists while funneling money back to CEOs and their Wall Street investors. “Gravity Falls” creator Alex Hirsch gave a blunt summary of this view in response to the Coca-Cola ad on X/Twitter, suggesting Coca-Cola loves the color red because “it’s made from the blood of out-of-work artists.” Ouch.
For those who work as creators, obviously, the threat of losing work to AI is already real. Some human animation team would’ve been hired to create Coca-Cola’s ad, if those machines hadn’t done it. But that’s not to undersell the copyright theft issue, either! After all, this year’s AI Coke ad isn’t even an original creation. It’s a remake of that 1990s commercial, which someone at the time had to think up and produce, and who almost certainly did not receive any kind of compensation this time around. AI legitimately can’t create its own imagery from scratch. It’s just a remix of what has already been produced before.
And AI companies aren’t paying the writers and artists from the last century from whom they’re borrowing for all of these ideas. Just this week, The Atlantic reported that AI software from Apple, Anthropic, Meda, Nvidia, Salesforce and Bloomberg – the top companies in the entire field – had been trained on data sets including 53,000 movie screenplays, and more than 85,000 TV scripts. That includes “The Godfather,” “The Simpsons,” “Seinfeld,” “Twin Peaks,” “The Wire,” “Breaking Bad,” “The Sopranos,” and even “Alf.” Not just scripted shows, either. The AI data sets even sucked up dialogue from Golden Globe and Oscar broadcasts.
These companies stole their screenplay data in the exact same fashion that other companies extracted YouTube videos for their training data: through captions and subtitles. In this case, the full library of a website called OpenSubtitles was compiled into a public data set called The Pile. Yes, that’s the SAME EXACT DATASET that was discovered to contain stolen, copyrighted YouTube transcripts back in July.
It’s not particularly challenging to make the case that AI is bad. I’ve been doing it in this column for over a year now. But it’s not the sort of thing that anyone can really snap their fingers and make go away. According to Menlo Ventures, companies have spent a collective $13.8 billion so far this year on AI, a 500% surge over 2023.
So creators face a divided landscape, in which no one is really happy with AI, but also it’s clearly not going away entirely, and a lot of very, powerful people have a strong incentive to make it happen. But it remains in an extended holding pattern. The only certainty about its future seems to be uncertainty.
Social Media Today reports this week that ByteDance, the Chinese company behind TikTok (at least for now), started developing a new project – called “V” – which can generate personalized AI avatars for influencers based on their “personality and thinking.” Theoretically, this would allow an influencer to be always online, able to interact one-on-one with fans in a way that was both authentic and automated. (Meta and Snapchat are working on similar technology to allow creators to build virtual versions of themselves.)
Will this perfectly or even semi-accurately recreate the actual experience of speaking live to your favorite influencer? Based on the recent applications of the technology – not to mention the waxy-skinned, dead-eyed stares of the actors in that Coca-Cola ad – probably not, at least for a while. But it may not even matter. OnlyFans gross payments surged by 19% in 2023, for a total of $6.63 billion on the year, and it’s widely known that fans are mostly interacting with imitators (or “chatters”) just pretending to be the sexy personalities behind the accounts. At this late date, OnlyFans users are either purposefully going along with the deception, or purposefully not taking the time to do any investigations for themselves. Same difference.
The same may be true of influencers and their audiences. Perhaps interacting with an AI avatar that seems vaguely like your favorite online personality will be deemed “good enough” to earn a profit, thus providing creators (or select creators, at least, with a particularly rabid fanbase) with a vital new revenue stream.
PRWeek published some new data looking into the phenomenon of “burnout” among professional creators, and whether or not it’s being impacted by AI technology. And the results are just as muddled and indistinct as all these other indicators. 71% of respondents said, at least so far, AI has had no impact on their feelings of burnout. That’s probably because most of these creators have not even tried to use AI apps yet. Only 19% of respondents from the US reported any experience working with AI. That jumps to 32% when you include everyone around the globe.
So even after 2 years, everything about AI and its impact on the entertainment and media landscape feels uncertain, and theoretical. Perhaps we’re waiting for the next major innovation that will move forward this technology and demand more immediate relevance. Perhaps the endlessly optimistic hypesters who insist that, 5 years from now, all of this discussion will sound quaint, and we just require more patience, have been right all along?
Perhaps, after sinking so many billions of dollars in venture capital into these firms, the elite establishment has already decided that AI is the future, and they’re simply waiting for the rest of us to diligently fall in line? I leave it you, gentle reader, to decide for yourself.
