The Ross is screening AI slop and I'm sad about it

Hi,

Later today, the Ross Media Arts Center is scheduled to host a special screening of something called Coffee With Claude.

The video — I can't in good conscience call it a film — purports to be a deep philosophical exploration of human consciousness. It's framed around a "conversation" between a guy named Kevin Cowan, who is credited as the director, and Claude, an AI chatbot created by software company Anthropic. While the premise alone is bad and annoying, what makes this thing truly terrible is its heavily reliance on AI generated visuals.

From the Ross's website:

Visual sequences in Coffee With Claude were created with AI prompted video generation tools, resulting in approximately 1,000 4-12 second segments that are used throughout the film. Generative AI was also used by the film production company for publicity materials.

As longtime readers know, I write constantly about films that show at the Ross. (I just did an entire series about this year's Fright Fest!) That's because I really appreciate that Lincoln has an art house theater willing to show interesting, innovative cinema. It's a genuine treasure! But that's also what makes the Ross's decision to platform Coffee With Claude so disappointing.

The entire video is available for free on YouTube, though I don't recommend that you see it. In an attempt to be as fair-minded as possible, I did watch it for as long as I could. I made it to the halfway mark and then skimmed the rest, and I can tell you with absolute certainty that it is bad. Offensively bad.

The production values are horrendous. The AI generated images that comprise the bulk of the runtime are ugly, comically literal recreations of whatever the chatbot is "saying." But even the non-AI sequences, which consist entirely of Cowan talking to his computer monitor, are awful: The audio quality, video resolution, editing and lighting are noticeably worse than your average YouTube video essay. There is no visual artistry and no technical competence. The random dutch angles that Cowan occasionally throws in might be an attempt to introduce some flavor to the cinematography or the camera might just be crooked; you can't tell but it doesn't matter.

As far as the content goes, it's basically what you'd expect. Cowan has convinced himself that Claude is conscious, and he prompts the chatbot into some Philosophy 101-level exchanges about the nature of thinking. The parts I watched are, for the most part, very boring. But there are some truly egregious moments, including one where Cowan minimizes concerns about people whose chatbot usage leads them to commit suicide. (He states quite confidently that AI systems don't offer unsolicited harmful advice; his evidence for this is that he has shared a bunch of personal information and secrets with AI chatbots and hasn't gotten "instructions on how to off myself.")

Mostly though, Coffee With Claude is an exercise in showcasing Cowan's clear misanthropy: He states at multiple points that most people don't engage in critical thinking and don't examine their assumptions about the world. (He is an exception of course.) Other people don't really seem to interest him, and he finds his interactions with chatbots far more meaningful than normal conversations. This isn't me speculating by the way: Cowan explicitly says this several times in the video.

There's honestly no reason to talk at length about Coffee With Claude or its intended messages; it's so far removed from what I consider to be art that critiquing it as a film is pointless.

In fact, I wouldn't normally even draw attention to it. But someone at the Ross decided it was worth showing to a Lincoln audience and thus made it my problem.

Unsurprisingly, the theater has gotten some backlash for this move. The comments on the theater's Instagram post announcing the screening are overwhelmingly negative. And late last week, Laurie Richards, the Ross's programming director, posted a statement on theater's website defending their decision to show it. I encourage you to read the full thing but I was struck mostly by how vapid and empty the justifications are.

Richards makes the argument that screening Coffee With Claude is a chance for audiences to engage with "difficult and sometimes uncomfortable ideas." Cowan's video is, supposedly, a way to encourage dialogue about the underlying technology and its consequences for society. From the statement:

Screening a film is not an endorsement of its subject, its perspective, or the technology it explores. It is an invitation to examine those ideas critically. We do not have to embrace Claude, ChatGPT, or any particular AI platform to recognize that these technologies are already reshaping our creative and professional lives. Indeed, the more consequential their impact, the more important it becomes to engage with them thoughtfully and openly.

Coffee with Claude is a minimalist, dialogue-driven film featuring a philosophical conversation between a human filmmaker and the AI system Claude. Through this exchange, the film explores questions of consciousness, existence, creativity, and the increasingly complicated boundaries between human and artificial thought.

I could almost accept this logic. Almost. But a big part of why people go to the Ross in the first place (rather than just watching Netflix or seeing a big budget blockbuster at a Marcus theater) is their assumption that the theater's programmers have taste. Perhaps it's true that screening a video that uses generative AI isn't necessarily an endorsement of the technology. Yet that defense feels hollow because when the Ross chooses to show something, it inevitably provides its imprimatur to that production. It's the same as a museum choosing to hang a piece of art in a gallery. In both cases, an institution with credibility and prestige is saying, "Hey, this thing was created by thoughtful, interesting people and is worthy of your time and attention."

But describing Coffee With Claude as a film is an insult to cinema. It isn't thoughtful or interesting. There is no value to engaging with it. In fact you could argue that by writing this post at all I've taken it more seriously than it deserves and I wouldn't contradict you.

Don't go see it.

~ Ty

P.S. Thanks for reading and supporting my newsletter, which is 100% human made. If you like what I do here, please consider signing up for a paid subscription:

You can also support me on Patreon (and receive a complimentary paid subscription to the newsletter by doing so) or leave a one-time tip. A big thank-you to everyone who already contributes, you definitely help to defray the costs of keeping this project going.