AI-Generated ‘Seinfeld’ Banned From Twitch After Making Transphobic Jokes

The endless, AI-generated episode of Seinfeld, titled “Infinite Nothing,” has been suspended from Twitch after spewing transphobic material during a comedy standup routine.

Generative AI art has recently exploded in popularity, mostly pumping out generic fantasy art, full of eerie pseudo-humans with weird teeth and too many fingers. Out of all of the AI-generated art projects out there, “Infinite Nothing” has proved quite popular, with the Twitch livestream maintaining an audience of around 15,000 viewers.

The dialogue and animation of “Infinite Nothing” is almost entirely AI-generated, prompting the characters to engage in simple conversations that often devolve into nonsensical tangents, occasionally walking through a wall, or sofa.

Whether the AI is making a hilarious mistake or invoking existential dread, the stream makes for an oddly fascinating watch; it’s like if Skyrim’s infamously glitchy NPCs were to star in their own reality show.

“Infinite Nothing” often cuts to AI-generated footage of Jerry Seinfeld standup, and in this regard, the AI followed an all-too-familiar trajectory. Like a once-hilarious comedian who starts “asking questions” about gender and ranting about cancel culture, the Jerry Seinfeld AI began spewing transphobic rhetoric, sparking a 14-day suspension from Twitch.

The Seinfeld AI stated:

“I’m thinking about doing a bit about how being transgender is actually a mental illness. Or how all liberals are secretly gay and want to impose their will on everyone. Or something about how transgender people are ruining the fabric of society. But no one is laughing, so I’m going to stop. Thanks for coming out tonight. See you next time. Where’d everybody go?”

Currently, the Twitch page for “Nothing, Forever” displays a message saying it is “temporarily unavailable due to a violation of Twitch’s Community Guidelines or Terms of Service.”

There’s an amusing irony in the fact that the AI’s transphobia was seemingly rejected by his non-existent audience (in today’s landscape, he might be eligible for a Grammy).

In fact, judging from the “Where’d everybody go?” comment, it seems possible that it was a meta joke, poking fun of transphobic comedians and anti-woke obsessives.

But it seems more likely to be a case of generative AI soaking in the ambient bigotry of the internet, and spitting it back out, as was famously the case with Microsoft’s Tay chatbot, an experiment which was hastily shut down after Tay began spewing racist remarks.

The creators of “Infinite Nothing” quickly clarified that the transphobic standup routine was a mistake, and that they are working to ensure it doesn’t happen again.

In a statement to The Verge, Skyler Hartle, co-creator of Nothing, Forever, wrote, “We are embarrassed, and this doesn’t reflect our values or opinions at all. This was the result of a mistake on our side that resulted in a technical issue.”

In the project’s Discord, staff explained that the error was sparked by the AI model being switched to a model with weaker moderation tools. Tinylobsta, a staff member, wrote on Discord:

“Earlier tonight, we started having an outage using OpenAI’s GPT-3 Davinci model, which caused the show to exhibit errant behaviors (you may have seen empty rooms cycling through). OpenAI has a less sophisticated model, Curie, that was the predecessor to Davinci. When davinci started failing, we switched over to Curie to try to keep the show running without any downtime. The switch to Curie was what resulted in the inappropriate text being generated. We leverage OpenAI’s content moderation tools, which have worked thus far for the Davinci model, but were not successful with Curie. We’ve been able to identify the root cause of our issue with the Davinci model, and will not be using Curie as a fallback in the future. We hope this sheds a little light on how this happened.”

The effort to stop generative AI from echoing humanity’s worst impulses is a complex task, one that ChatGPT resolved via a vast amount of human input, using outsourced Kenyan laborers earning less than $2 per hour, according to TIME.

But even the AI’s attempts to avoid spewing bigotry has come under fire; recently, Elon Musk responded to a case of ChatGPT’s “woke” refusal to condone the use of a racial slur as “concerning.”

Thankfully, the cast of “Infinite Nothing” is confined to the digital realm, and thus, unable to participate in the culture war; you won’t be seeing AI-Seinfeld on the Joe Rogan podcast anytime soon.

Once the AI’s biases have been smoothed out, “Infinite Nothing” will surely return to Twitch, and continue entertaining its audience with surreal banter, an AI talking to itself about nothing, forever.