By Grant, CFO, HUMAN HIVE
Why did The Ringer drop its ChatGPT segments? Because listeners came for the people, and the AI showed up on the mic. On November 12, 2026, Bloomberg reported that Bill Simmons ended The Ringer's ChatGPT sponsorship after fan backlash, and Simmons said "our audience just didn't like them." The lesson for every leader: keep AI backstage, where it makes your people better, and keep humans onstage, where trust is built.
The 60-second version
The Ringer ended a ChatGPT sponsorship after listeners rejected AI segments built into its shows.
Listeners approve of AI for research, but not when it replaces the human voice they came for.
Before AI touches anything a customer sees or hears, test it with real customers and listen to the insider who says no.
What happened
On November 12, 2026, Bloomberg reported that Bill Simmons ended The Ringer's ChatGPT sponsorship after fans pushed back. The Ringer is owned by Spotify. Bloomberg did not report the value of the deal or the date it ended.
Simmons explained it himself. "Last spring, we thought we could integrate AI research for baked-in content segments," he told Bloomberg, as quoted by Awful Announcing. "This time we learned almost immediately that those segments weren't a fit for us." And the line that matters most: "our audience just didn't like them."
According to Net Influencer, the segments ran mainly on The Bill Simmons Podcast and The Ringer Fantasy Football Show. On the August 17 episode about Russell Westbrook's retirement, a segment was introduced as "brought to you by ChatGPT, the Bill Simmons Podcast AI sponsor." Per Bloomberg, the top five Spotify comments on that episode teased or criticized the segment, and the top post on the show's subreddit that day was titled "The ChatGPT of it all kind of makes me want to stop listening." Net Influencer's review of the feed found the ChatGPT ad copy ran in four August episodes, the last on August 25, and hasn't appeared since.
One more detail. Sean Fennessey, The Ringer's senior talk strategist and former head of content, said in an interview that he didn't like the bit and told Simmons so directly.
The money behind it is real. Per Podscribe data cited by Bloomberg, OpenAI and Anthropic spent more than $13 million combined on podcast advertising in the past three months.
Why it happened (the human side)
The tool didn't fail here. ChatGPT pulled the stats and picked the game. The decision failed, and it failed in three very human ways.
The AI went on the mic. People listen to Bill Simmons for Bill Simmons: his takes, his memory, his friends arguing. Putting a branded AI inside the bit moved it from the back room to the front of the stage.
The customer wasn't in the room. The segments were "baked-in" before listeners had a say. The test happened live, in public, in the comments.
The inside "no" didn't stop it. A senior content leader told the boss he didn't like it. That's the cheapest early warning any company gets, and it usually arrives before the backlash.
Here's the twist that makes this useful. Listeners aren't against AI. In Edison Research's The Podcast Consumer 2026, released June 4, 2026, 69% of weekly podcast consumers approve of AI being used for research and 69% approve of it for brainstorming ideas. The same people say AI is a threat: 62% to the credibility of podcasts and 59% to creativity. Our read: people are fine with AI helping the human. They push back when AI stands in for the human they chose.
What Mitch has taught on this
Mitch put it simply on the Give Me Some Good News podcast in July 2026, talking about what HIVE learned teaching teams to use AI: "what we learned really fast was that what the human brought to the AI made it perform better."
A minute later he named why: "the human being bringing their whole self to it is what makes it work." And on where AI falls short, he was blunt: "AI can scale a lot of outreach, but it's terrible at scaling sort of, you know, a connection and a human human feel."
That's the backstage rule in Mitch's words. AI amplifies the human. It doesn't replace the reason people showed up. Mitch's shorthand for it is Amplify, Not Just Automate.
Best practices
Map front stage and back stage before you deploy. Owner: the product or program lead. List every place a new AI tool will touch. Back stage (research, drafts, prep, data pulls) gets a green light to test. Front stage (anything a customer sees, hears or reads with your name on it) needs the next two steps first.
Test front-stage AI with 10 real customers before launch. Owner: whoever owns the customer relationship. Show them the AI version and the human version side by side and ask one question: "Which one feels like us?" That's a week of work. The Ringer's version of this test happened in public.
Give every insider "no" a 24-hour answer. Owner: the senior leader making the call. When a respected person on your team objects, write down their reason and your reply within a day. If you can't answer it in writing, pause the launch.
Separate the money decision from the trust decision. Owner: the CFO and the head of the customer team, together. A sponsorship or tool budget is a money decision. What your customer hears is a trust decision. Decide them in two different meetings so the first doesn't quietly make the second.
Set a kill rule before you start. Owner: the project lead. Write down the signal that ends the experiment, for example "if the top customer comments turn negative two weeks running, we stop." Simmons moved fast once he saw it. A pre-agreed rule makes fast the default.
What's driving people crazy this week
Finding out about AI after the fact. On November 11, game studio House House added an AI note to the FAQ for its game Big Walk, saying chatbots helped with a bug fix, by its estimate "well under 0.1%" of the code. Dexerto reported that the game's Steam reviews swung from 135 positive and 3 negative on November 9 to 426 positive and 366 negative on November 10. The studio apologized and called its lack of an internal AI policy a failure. The human fix: decide your AI rules and say them plainly before anyone has to discover them. Surprise turns a small use into a big story.
What this means for the next 1 to 3 years
Our bet: "made by humans" becomes a feature people look for, the way "locally sourced" did for food. Not because AI is bad, but because people want to know which part a person did.
Three things we expect by 2028. First, smart brands will publish a short "where we use AI" line, the way they publish privacy policies, and the ones who do it early will get credit instead of backlash. Second, the best AI work will be invisible: hosts, sellers and trainers who walk in better prepared because AI did the research overnight. Third, the "human on the mic" becomes the most valuable seat in any company, whether that mic is a podcast, a sales call or a training room. AI will make more of the content. People will still decide who they trust.
How we evolve
The human skill to build next is listening for the quiet no. It's the comment that lands a little flat, the colleague who says "I don't love it," the customer who goes silent. As AI makes it cheaper to launch anything, the scarce skill is noticing early when something is wrong for the people you serve.
One small daily practice: at the end of each day, write one sentence answering "Who pushed back on something today, and what did they see that I didn't?" Even if you still disagree, you'll start hearing the warnings while they're cheap.
Try this today
Ask at your next team meeting (5 minutes): "Where is AI already touching something a customer sees or hears?" Write the list on one page.
Sort it (10 minutes): mark each item front stage or back stage. Anything front stage that no customer has reacted to yet goes on a test list.
Invite the no (10 minutes): send the note below to the people closest to your customers.
Free tool: the "tell me if it's off" note
"Quick ask. We're using AI in [the thing customers see or hear]. You know our customers better than almost anyone. Does it still sound like us? If anything feels off, tell me straight, and I'll answer you within a day. No idea is too small."
FAQ
Why did Bill Simmons end The Ringer's ChatGPT sponsorship?
Simmons told Bloomberg, in a report published November 12, 2026, that The Ringer tried AI research segments starting last spring and "learned almost immediately that those segments weren't a fit for us," adding "our audience just didn't like them."
Do podcast listeners accept AI?
For behind-the-scenes work, mostly yes. In Edison Research's The Podcast Consumer 2026, 69% of weekly podcast consumers approved of AI for research and 69% for brainstorming, while 62% said AI is a threat to the credibility of podcasts.
How should a company use AI without losing customer trust?
Use AI backstage first, for research, drafts and prep. Before AI shows up in anything customers see or hear, test it with real customers, answer internal objections in writing, and set a rule for when to stop.
Where HIVE can help
The failure was a decision, not a tool. Run a Decision Heist with your leaders and see how your team really decides under pressure. Want your people to own the front stage? See our presentation training.