'Searching for SmarterChild' Wins at Ji.hlava, Asks Why We Abuse Chatbots

A documentary about why humans insult chatbots — and whether they should apologize — has won the New Visions Award for the most promising U.S. project at the Ji.hlava Documentary Film Festival this week. The film is "Searching for SmarterChild," co-directed by Zan Gillies and Pulitzer-winning journalist Lindsey Sitz, and produced by Erin Williams-Weir.

Its subject is a piece of internet archaeology. SmarterChild was one of the most popular early chatbots in North America two decades ago, reaching millions of households subscribed to AOL, Yahoo and other platforms. Around 3 million people exchanged text messages with it — many of them, says Gillies, markedly rude.

"We went out to San Francisco and interviewed people on the street to see if they remembered SmarterChild and to see if they had said something they wanted to apologize for," Gillies says. "One guy specifically said, 'Oh yeah — I was so mean to that bot.'"

What does the film actually ask?

At the project's core, Gillies says, is the question: "What does our relationship with artificial intelligence reveal about ourselves?" The technology has changed since the early internet, but the underlying questions haven't, he argues.

"We've got this half-human sounding thing and when we communicate with it what does that do to our brains? What is our relationship to these things? Are they our friends? Are they our therapists, or assistants or lovers or whatever?"

The pattern isn't uniquely American. The co-directors found that early chatbot experiences elsewhere mirrored the North American one, with users in Korea and beyond also unloading insults on text platforms. "A lot of us just have this initial tendency to, uh, be mean to robots," Gillies says.

A confession in the archives

The project took shape about two years ago, when Gillies and Sitz realized they had each used SmarterChild as young people — and each had been abusive to it.

The typical opening went like this: the bot would ask the user's favorite hobby and get a hostile answer such as "Setting things on fire." It would politely suggest an alternate activity, then be told something like "Setting YOU on fire."

Those responses, Gillies discovered, were literally his own words. He tracked down the inventors and engineers behind SmarterChild, who still hold digital records of every chat session ever conducted.

The filmmakers have so far interviewed members of both the West Coast team and the East Coast team that wrote SmarterChild's response language, before the bot was sold to Microsoft and shut down.

How Silicon Valley handled the abuse

The engineers' side of the story is its own chapter. After SmarterChild's initial success, the Silicon Valley team was surprised by how abusive users were, Gillies says, and convened emergency meetings to deal with it.

Their fix: a feature letting SmarterChild recognize rude language and respond with warnings — and even a degree of snark. If the abuse continued, the bot would go into temporary shutdown until the user apologized.

One detail from the research genuinely startled Gillies. The programmers had built a rating system for calculating the relative weight of insults. "I thought that was the funniest thing," he says.

What happens next

The film isn't finished. Gillies and Sitz have raised about $18,000 through crowdfunding so far, which they expect will cover roughly a fifth of the final budget.

They plan to interview more participants and experts on human-AI interaction, including a noted expert on apologies who works in conflict resolution. They have also tracked down Lauren Kunze, CEO of the chatbot company Pandorabots, who has researched which parts of the human brain are triggered by AI interaction.

Kunze is one of several on-camera figures slated for the final cut. The filmmakers also want to include a sequence in which a rebuilt SmarterChild bot converses once again — which means audiences may yet get the chance to watch an old insult meet an old machine.