Why do brands keep humiliating themselves with AI slop?
Major companies keep embarrassing themselves by publishing nauseating AI slop again and again, never learning. Why?
The Washington Post December 10 released a new AI-powered personalized podcast feature unimaginatively called “Your Personal Podcast.” The Post’s goal was to attract younger users to its news brand by offering pick-your-own-format news podcasts. When you set up the app, users can choose topics, hosts, and duration.
The inevitable push-back started the very next day.
Online critics pointed out that the podcasts were packed with errors, mispronunciations, and made-up quotes attributed to people who never said them.
Worse for the reputation of the Post (owned by bookish bookseller-turned-supervillain Jeff Bezos) the AI misconstrued a source’s quote as the official position of the newspaper.
Management knew the product was garbage. During pre-launch testing, all the errors and hallucinations were easy to spot. But they launched anyway, based on the old Silicon Valley approach of correcting iteratively over time with an already-shipping product.
Employees at The Washington Post organized an internal protest. Staffers expressed their distress and anger through Slack messages. It was their reputations being harmed by a publication whose main currency is credibility.
Shitty AI-generated podcasts are trending. The podcasting industry is being “flooded” with thousands of AI-generated podcasts. To sound more human, these AI voices now include speech tics. They use “ums,” “uhs,” and prolonged pauses to simulate natural conversation. The text-to-speech tech is so good that some listeners can’t tell the difference.
Platforms like NoteGPT and Google’s Audio Overview (launched late 2024 in NotebookLM) have democratized podcast creation.
Studios are releasing “hundreds of thousands” of AI-generated episodes. The volume is so high that it’s “swamping” the industry.
Inception Point AI, an L.A. startup, churns out some 3,000 episodes weekly with a team of eight employees (total output: 200,000 episodes).
There are currently at least 175,000 AI-created episodes on major platforms like Spotify and Apple Podcasts.
Everyone knows the main benefit of AI-generated content is that it’s cheap. So nobody’s impressed by it, especially when it’s filled with errors.
Why do companies keep doing this knowing that others get punished by the public?
McDonald’s faced backlash this month after releasing an AI-generated Christmas commercial titled “The Worst Time of the Year.” Critics were not lovin’ it, and slammed the video as “disjointed,” “uncanny,” and “stupefying.” McDonald’s had to McDisable comments, then McRemove the slop from public view on December 11.
McDonald’s isn’t the only obesogenic junk food giant slinging slop in ads. In November 2024 and November 2025 Coca-Cola used AI to recreate its old “Holidays Are Coming” commercials. The ads were slammed for being “soulless,” “dystopian,” and “devoid of any actual creativity.” Car blog Jalopnik pointed out that the truck in the ad had 10 different axle configurations during the 60-second commercial.
Earlier this month, the luxury fashion house Valentino faced criticism for a “lazy” and “cheap” campaign that used AI to generate handbag images, and also to lean into AI slop as an aesthetic.
There are dozens of other examples too banal to mention.
The "McRemove" line took me out. 😂
But you nailed the core issue with the WaPo example. They are applying the "ship it and patch it" mindset to an industry where trust is the only product. You can patch buggy code, but you can’t patch a hallucinated quote. Treating credibility like a software beta is wild.
What's crazy to me is that some of these companies have loads of cash yet are accepting the worst, cheap AI outputs as ads.