Dafuq were you thinking Google? AI Ad for Olympics is just plain weird
The Tech industry continues to grasp for an AI future that isn't weird, and they flail ... again.
One of the newspapers I subscribe to is The Atlantic. This morning I was doing my usual scrolling through the app, and I stumbled across this article: “Google Wins the Gold Medal for Worst Olympic Ad” (gifted link - valid for 2 weeks), that title grabs my attention, and upon reading the article, I think to myself that they need to fire their marketing team.
The gist of the ad is:
Google is running a new commercial during the Olympics. It’s about a cute little girl—she’s a runner, and she loves Team USA’s Sydney McLaughlin-Levrone, a world-record-holding track star who won two Olympic gold medals in 2021. The little girl wants to write her a letter. So Dad fires up an AI chatbot.
Sounds endearing, right? Well, it did until the last sentence.
Look, the way young people look up to their idols, who want to follow in their footsteps is a normal phenomenon. Like letters to Santa, they are highlights in a young person’s life. My wife’s niece was a huge fan of Kerri Walsh (Jennings) a volleyball player who won gold in the Olympics, and her niece did follow the path, ultimately playing for Northwestern on a scholarship.
So, this is a real phenomenon.
But the charm of it, for both sides, is the youngster writing a heartfelt, probably in poor penmanship, and with grammatical and spelling consistent with their age and education level.
Instead of a negative, that shows humanity.
But there’s no way that Google can market that. I mean they are spending billions of dollars, and consuming YUGE amount of electricity, up 13% as reported in their recent earnings report, to fuel their generative AI with their Gemini chatbot built to spar with OpenAI’s ChatGPT.
So, some marketing dweeb pitched ideas for commercials to air during the Olympics to promote their Gemini product.
And the one that broke through was using Bard to write the letter from his daughter to her idol who is competing.
“Help my daughter write a letter telling Sydney McLaughlin-Levrone how inspiring she is,” he asks Google’s Gemini. He instructs it to add a line about how his daughter plans to break McLaughlin-Levrone’s world record one day (and to be sure to include the phrase sorry, not sorry.)
Really? How is that going to inspire your daughter? How is Sydney supposed to react to a generated AI message from a young girl? Where’s the emotion, the connection, the humanity?
C’mon fam, don’t do that shit.
The Atlantic agrees with me, the next snippet from the article continues:
The ad never shows the final letter in full, just pans over snippets of it. The whole thing is supposed to be endearing, demonstrating to viewers how AI can help forge human connection and facilitate creativity.
But come on: Nothing about this ad makes sense.
Isn’t what makes a letter like this cute the fact that it is written by a child? Shouldn’t a young person get to explore their feelings and then authentically relay them? And what about McLaughlin-Levrone? Will she be able to tell that the letter was written by AI? How would she feel about that? Would she send back her own AI-composed message, thanking the child for taking the time to write to her?
This reminds me of a conversation that I had with a colleague at work. After a round of layoffs, one of her impacted employees asked for a LinkedIn recommendation, and being short of time, she just used ChatGPT to write it1. It looked OK, so she copy-pasta’d it. But at least she felt a little bit of shame.
Side note: HR systems are already using AI/ML tools to scan, process and confirm applicant’s qualifications, and yes that includes hoovering up shit from LinkedIn, so you have AI talking to AI to sort applicants so hiring managers only get “good” resumes to read. This is fine.
The whole thing is bleak. It takes the feel-good cliché of a child getting to interact with their idol and squishes a multimillion-dollar large language model between them. Google is pitching a world in which even the most personal interactions are mediated by computers. The company may make bold claims about AI’s capabilities to radically advance civilization. But it can’t escape the reality that it’s co-opting the hopeful aesthetics of the Olympics, which are meant to celebrate human accomplishments, in order to promote a digital technology that can be used to undermine human labor. (emphasis mine)
That is the central theme in my mind. Do we really want to outsource our intelligence to a machine? I am sure there are plenty of people who are “hell yeah”, but to me, a lifelong learner and student, living a life of wonder and discovery, this is hell.
It is already bad enough that we all have this device in our pocket where we can access literally any piece of information in the world with our fingers, but to take the human component up to a level of abstraction where there is no longer any connection to the knowledge they consume.
Naturally, the ad generates cringe:
The reaction so far has not been positive. The author Will Leitch said the ad “takes a little chunk out of my soul every time I see it.” The professor and media personality Shelly Palmer wrote that it makes him “want to scream.” On YouTube, where Google posted the ad four days ago, comments are turned off—a step that the company does not typically take on its videos, and one that suggests concern about a backlash.
I am certain that there is a lot more of that out there. And this tracks, because this week when OpenAI launched their SearchGPT, and it sucks. From that article:
n a prerecorded demonstration video accompanying the announcement, a mock user types music festivals in boone north carolina in august into the SearchGPT interface. The tool then pulls up a list of festivals that it states are taking place in Boone this August, the first being An Appalachian Summer Festival, which according to the tool is hosting a series of arts events from July 29 to August 16 of this year. Someone in Boone hoping to buy tickets to one of those concerts, however, would run into trouble. In fact, the festival started on June 29 and will have its final concert on July 27. Instead, July 29–August 16 are the dates for which the festival’s box office will be officially closed. (I confirmed these dates with the festival’s box office.)
Yeah, not great that it can’t connect the dots with publicly available data. Sorta like the reasonable sounding hallucinations that are made up, and now it is for items a user is searching for.
Oops.
Back to the original premise though, as the Tech-bro Oligarchs fight amongst themselves to up their AI game, they are also struggling to find real uses that aren’t cheating on homework, or making images of Trump as a jacked Jesus with washboard abs (and too many fingers).
But in the real world, the tendency of these models to make shit up limits their utility in real use cases. There’s the story of the lawyer who submitted a filing where ChatGPT made up non-existent citations, and the judge nailed him for it, or when a co worker was using it to research devices for a network design, it spit out fake product SKU2’s that didn’t exist and thus couldn’t be ordered.
Oops.
But that doesn’t mean that greedy executives and middle managers aren’t using these generative chatbots to replace the cogs in the machines of their organizations. Because chatbots can crank out generic ad copy, or brochure blurbs that sound reasonable (but often are very flat when viewed from a distance). Many mediocre bosses will be fine with that, but the implication is that a LOT of the entry to middle class jobs and roles will be vaporized, and in a generation, there will be essentially the wealthy 5%, and the poor 95% with little in between.
Google’s marketing team has a tough job right now. The company has aggressively pivoted to a technology that may be dazzling, but that many people remain skeptical of. AI will revolutionize everything, boosters say, but it’s still unclear exactly how. Wall Street is starting to wonder whether investments in the technology will actually pay off. To the extent that generative AI is present in everyday life, it’s not always on the best terms: The technology has arguably degraded once-reliable search engines, plundered human creativity, and taken jobs.
All of which is to say, the reality is far from the sunshine and jump ropes of the “Dear Sydney” ad. Perhaps that’s not unusual: For years, Big Tech’s marketing has relied on sweet montages of regular people using their tools to skirt the very real problems presented by their products. The likes of Meta, TikTok, and Apple may be able to get away with this framing, because their products do connect people at the end of the day, but generative AI is more about humans talking to a computer instead of one another. (Apple found itself in a similar situation earlier this year with an iPad ad that, accidentally or not, evoked AI’s ability to crush art with a machine; Apple quickly apologized and halted plans to run the commercial on TV.)
Personally, I think Google’s marketing team should be fired en masse and replaced with people who have a scintilla of common sense.
The truth is the current crop of Generative AI tools are cute toys. If you are skilled at asking questions, and can geek out with refinements, you can get something reasonable out of it. But is it a miracle? Not really, and are the inflated stock prices and valuations of OpenAI justified? Judging by the increasing stream of stories from the financial press that are asking questions about when the tens of billions of dollars of investment will begin delivering solid returns, as well as a recent pullback of Nvidia’s stock price, the answer is that Wall Street is adopting a “show me” attitude, and while not shunning the space, they are beginning to hold the companies’ feet to the fire.
Time to put up or shut up.
I will close with the last pull from the article:
Google appears to have misread the moment. The Olympics are supposed to be about humans accomplishing amazing feats in the physical world. While the ad was running this weekend, the American surfer Caroline Marks scored a close-to-perfect 9.43 points out of 10 after dropping into the barrel of a giant wave in Tahiti. The 17-year-old Canadian swimming prodigy Summer McIntosh won her first gold medal for the women’s 400-meter individual medley. And the legendary gymnast Simone Biles continued to defy the laws of physics despite an injury. These athletes are indeed inspiring. We don’t need a chatbot to tell them so.
I wholeheartedly agree!
For the record, every recommendation I have ever given, and will ever give on LinkedIn are words that I personally wrote, and it represents my real impressions of the person I am recommending.
A SKU is a “Stock Keeping Unit” that is a product with a part number that can be ordered.
It perhaps goes without saying that, as long as AI search engines can mistake something as simple as open and closed dates, it is essentially a complex toy, not a real tool!
This exactly reflects my limited experience with AI. I remain interested but I don’t feel it is revolutionary. It is similar to “revolutions” in programming languages. I have generally found that I can do anything in the FORTRAN that I learned in 1964 or its subsequent revisions that can be done in newer languages and those mostly represent someone’s preferences. The exception is for a language like PYTHON for which a highly involved user community has developed useful tools and instruction. Similarly, I have found that AI searches are ultimately no better than old style keyword searches provided one is adequately clever. I certainly don’t bow to AI for its writing ability or supposed superior understanding of complex subjects. Finally, I live to learn and will not surrender my intellect to a machine.