“ChatGPT Voice Grandmas” — AI privacy and the average consumer
This past week, I spent my evenings binge-watching “Chad Powers”. If you’re not familiar with the show, it’s about a college football team and their star player. Given the content of the show, my guess is that the audience skews male, probably between 24–50 years of age. I wasn’t able to confirm these stats, so that remains my guess. My focus, though, was not on the show itself — the show is great. It’s ridiculous in all the right ways and has just enough emotional fun to keep non-sports fans engaged. Rather my focus was on the advertising during the shows.
Ok, so keep in mind my guess for the audience skewing male and under 50. That matters. With the ads, there were the usual ones for DriveTime, AutoTrader, and a myriad of drugs with three syllable names ending in -vy or -zi/zy. I expect those every time I watch anything since the algorithm has identified us as of car-buying age and in need of all the different drugs that might currently be available. What I was dismayed to find were the ChatGPT and Gemini ads. Both of which were normalizing the use of these tools in general life.
You can see the ChatGPT ad here with grandmas talking to “Chatty” while doing normal life tasks. I was bothered by these ads because the grandmas were talking to the voice like it was a friend — it felt like normalized AI companionship in a way that I’m not sure I’m ready for.
The conversation went back and forth between the grandmas and the voice, like the conversation that the ladies might have in a social circle. Vimalan (2026) notes that “[t]he model can acknowledge users with cues such as “mhmm” or “yeah”, pause while they gather their thoughts or continue listening without interrupting, creating a more natural conversational flow.” While some people might find this comforting, it feels like we are standing on a slippery slope into AI companions so that we can ignore people while also feeling ok about the fact that people still have “someone” to talk with. I’m still wrapping my head around the technology, but I’m not really on board yet.
The other ad that kept coming up was a man, probably in his early 30s, using Gemini AI to search for subscriptions that he had forgotten about. Gemini then identified these subscriptions so that the man could cancel them. Think RocketMoney, but instead you give an AI tool access to your email (granted, Gemini is already in Gmail, so that’s a lower lift). Sadly, this ad is appearing elusive enough that I can’t find it right now — even Gemini couldn’t find it. This ad normalizes allowing an AI tool access to your accounts in a way that seems to be priming the public for agentic AI to slide into our daily lives.
Both of these ads appear innocuous on the surface. Grandmas having a chat with a friendly companion and a man using an AI tool to save money. Both ads even make AI seem like useful and safe technology. By design, we are made to feel that these uses are normative for daily life. In order for the companies building AI tools to eventually make money on their gambles, they need to ensure that their tools are everywhere and in everything. They need people to become dependent on the faster, close enough answers and simple companionship that these tools project. Use cases like the ones above make sense to the average American looking for an easier way to accomplish a task.
Get Laura Dumin’s stories in your inbox
Join Medium for free to get updates from this writer.
But what do we give up as these tools become more embedded in our daily tools and tasks?
What does privacy count for in an era when it’s so easy to link all of our logins to Facebook or Gmail? When we can allow our phones and computers to access accounts across platforms so that we have fewer logins to remember? When we can save our credit cards to our computers and accounts so that paying for things is no more difficult than clicking a button? These are the questions that I’m grappling with right now.
For people just starting to consider what we give up for convenience, may I suggest reading Privacy is Power: Why and How You Should Take Back Control of Your Data
by Carissa Véliz. Véliz writes about privacy in ways that will keep you up at night, but maybe that’s not a bad thing. As we consider privacy from the basics of not allowing cookies to track you on the web to more advanced concerns of allowing agentic AI to have access to your accounts and complete tasks on your behalf, it makes sense to deeply consider what we are giving up in order to save a few seconds/minutes/dollars. I’m hoping to dive more deeply into this topic as I have time over the next few months. In the meantime, I’ll leave you with this thought:
Agentic AI is not inevitable. Just like chatbots and LLMs, we are the ones in the driver’s seat of personal consumption. We get to decide when and where these technologies make sense for our own use. While we can’t control what large companies choose to use AI for we can choose which companies we want to interact with and give our money and data to. Maybe that’s the first step in deciding what the future of AI actually looks like.
References
MediaPost Live. (2026, July 8). ChatGPT Voice Grandmas. YouTube. https://www.youtube.com/watch?v=BB3SWowrJsQ
Vimalan, D. (2026, July 13). OpenAI’s newest AI ambassador? Someone’s grandma. Marketing-Interactive. https://production.marketing-interactive.com/openai-s-newest-ai-ambassador-someone-s-grandma
