Can ChatGPT Recommend Movies? A Film Buff Put It to the Test

I told ChatGPT I enjoyed the 2013 film “Her,” whose protagonist develops a relationship with a virtual assistant. It spewed out a list of sci-fi titles like “Blade Runner 2049” and “Ex Machina.” “These movies,” it typed, “explore the relationship between humans and artificial intelligence, touching on themes such as consciousness, identity and the nature of existence.” (It gave no sign it saw the irony.)

Wei Xu, an interactive computing professor at the Georgia Institute of Technology, explained how ChatGPT managed to produce a list of legitimately comparable movies in seconds. The software, she said, is trained to spot patterns within a massive amount of text data—over 500 GBs—it scrapes off the internet. When sniffing out cinematic cousins to “Her,” it’s likely consulting sources like Reddit threads, IMDB forums, even “Best of” lists from editorial outlets. Traditional recommendation engines, said Dr. Xu, don’t have this access.

This presents issues for existing discovery platforms like Letterboxd, a social-networking site for discussing movies, and Likewise, a content-recommendation service that draws on AI and human curators. Letterboxd co-founder Matthew Buchanan told me he’s concerned by ChatGPT’s lack of transparency. To get the info it uses to make recommendations, it could be plagiarizing the work of Letterboxd users without providing credit. (OpenAI declined to comment for this article.)

For now, Mr. Buchanan says he’s taking solace in the fact that ChatGPT’s “anodyne” responses lack a human touch. I can’t help but agree. The humor and strangeness of Letterboxd reviews can leave me excited to watch particular movies. (In reference to the cinematographer of “Blade Runner 2049,” for instance, one Letterboxd reviewer wrote “I’’m pregnant and the father is Roger Deakins’ camera.) ChatGPTs responses are usually reasonable, but they rarely surprise.

And because the bot is only trained on data that predates September 2021, it has some blind spots, though I haven’t yet encountered these. In any case, the last time I settled in to watch a film, I knew exactly what to stream—I had to see what all the Roger Deakins fuss was about.

3 More AI-Augmented Apps

Tech companies are racing to incorporate ChatGPT-like capabilities into their own products. The following apps are free to download, but access to some of their features might cost you extra.

Simpler Searching

Neeva, a private search engine designed by

Sridhar Ramaswamy,

the former lead of Google’s ad and commerce division, uses AI to summarize the results of a traditional list of blue links, fetching one final answer to your query. No more sifting through a pile of obscure websites just to find the difference between baking powder and baking soda. (Expect more of this: Microsoft announced last week it was integrating the tech behind ChatGPT into its search engine Bing.)

Modern-day Memex

Mem uses AI to form contextual links between your emails, calendar appointments, memos and more so that you can find them all in one place. If you’re looking to attend a certain meeting, for example, Mem will create a page where you will find the joining link, plus any relevant notes you’ve prepared and related emails you’ve received and sent.

Easier Editing

You no longer need high-end software to edit videos, podcasts and voiceovers. Descript converts these into easily scannable transcripts, so that you can cut filler words or move scenes around. Used judiciously, this saves time, and sounds surprisingly professional.

The Wall Street Journal is not compensated by retailers listed in its articles as outlets for products. Listed retailers frequently are not the sole retail outlets.