Artificial Intelligence Is Already History, According to New Silicon Valley Exhibit
From medieval folklore to ChatGPT, the Computer History Museum's "Chatbots Decoded" traces the technology's path
It’s been just over two years since the release of ChatGPT made generative artificial intelligence (GenAI) part of the zeitgeist, but the nascent technology is already the subject of a new history exhibit.
With “Chatbots Decoded: Exploring A.I.,” Silicon Valley’s Computer History Museum hopes to help the public realize that despite the novelty of the technology, “there is a long history there, and it’s fulfilling a long dream,” says Marc Weber, the museum’s internet history curatorial director.
Through a series of artifact displays, interactive experiences and immersive video, “Chatbots Decoded” traces the journey of AI from its origins in the human psyche to its present-day ubiquity, and to an unwritten future that can still be influenced by an informed digital citizenry. “We wanted to create a fully immersive and interactive exhibit that not only traces the evolution of chatbots but also invites visitors to think critically about AI’s growing role in our lives,” says Kirsten Tashev, chief curatorial and exhibitions officer at the Computer History Museum.
AI and Anthropomorphization
AI’s role in the human experience began long before large language models made chatbots a practical tool. Considering this, the Computer History Museum’s exhibit tells a story of AI that includes literature, folklore and automata dating back to medieval times.
Moving into the 20th Century, the exhibit features artifacts from pioneering attempts at AI, including a physical, switch-based neural network from the 1950s. But it also includes artifacts whose relation to AI is more conceptual than technological, like the Chatty Cathy talking doll of the 1960s, the Pet Rock of the 1970s and another pseudo pet—the digital, portable Tamagotchi of the 1990s.
While these curiosities weren’t AI as we know it today, they demonstrate the power of anthropomorphization —the assignment of human qualities to non-humans—as one of the signature traits of the modern chatbot. “It doesn’t take much to get people to feel connected to something,” Weber says.
Chatbots Catalyze the AI Revolution
It’s that human-like connection that has brought AI to the surface after the technology spent decades behind the scenes, emerging into practical reality in fits and starts. This history, as covered in “Chatbots Decoded,” includes the mid-20th-century contributions of AI pioneers like Alan Turing, Frank Rosenblatt and Joseph Weizenbaum, early web-based chatbots and IBM’s Deep Blue, the AI system that famously defeated world chess champion Gary Kasparov and now has a home at the Computer History Museum.
“AI has done more than people realize in the past 60 years,” Weber says. “We use intelligent systems for buying stocks, automating financial programs, planning and logistics, spellcheckers. But we couldn’t talk to it.”
The curators at the Computer History Museum realized that limitation was no more when the first iteration of ChatGPT was unleashed in November 2022, astonishing the public with its ability to interact with humans in plain language. The revelation was “kind of like a dam bursting,” Weber says, and museum organizers fast-tracked the exhibit’s development to seize the moment.
Surviving ‘AI Winter’
The result of the museum’s quick action is a historical perspective on a still-unfolding phenomenon. While chatbots and AI might make people feel like they’re living in a sci-fi movie, the technology is actually decades late as far as the genre is concerned. For instance, “2001: A Space Odyssey” envisioned HAL 9000 as a sentient artificial intelligence navigating a spaceship through the cosmos at the turn of the 21st century. From that perspective, the artificial intelligence in existence today—not to mention space travel technology—is still behind schedule.
Over the past 50 years, poor PR caused a series of “AI winters,” stymying the technology’s development until the recent breakthrough of large language models, Weber explains. The first of those fallow periods came in the 1970s, when funders like the U.S. military’s Advanced Research Projects Agency (ARPA) became disenchanted at the technology’s failure to live up to hopes.
This underperformance gave AI a bad name, “and that’s why a lot of what’s actually happened in AI since then has not been called AI,” Weber says. Another “AI winter” arrived after IBM’s original Watson and Deep Blue proved to be relatively limited. Apple’s talking assistant, Siri, brought yet another wave of AI promise but was ultimately viewed as less than transformational.
So, technology that could fit the definition of AI has been gone by other names—spellcheck, optical character recognition, machine learning, automatic translation. Anything but AI. But now, AI is out in the open, and showpieces such as the humanoid, interactive, ChatGPT-based robot Ameca, created specifically for the “Chatbots Decoded” exhibit, can express themselves freely (or at least within the guardrails established by their creators).
The Unimagined Imagination of AI
While talking toys and sci-fi might have prepared humanity for AI to some degree, not everyone is taking its arrival in stride. “For the first time, we can actually talk to something that is not a person in this easy, conversational way. And it kind of breaks our brains,” Weber says.
Sci-fi didn’t envision the kind of creativity seen in today’s GenAI applications, he explains. “Chatbots today are creative, to the point that they threaten the human creative class,” Weber says. To illustrate, he points to the writers’ strike that brought Hollywood to a halt in 2023 as movie and TV writers pushed back against AI and other threats to their livelihood. As a testament to the strike’s historical significance, “Chatbots Decoded” features picket signs from the 148-day work stoppage.
Like the 1990s, But With a Twist
While anxieties about AI replacing human creativity persist, the prevailing attitude among investors and companies is “very similar to the way it was 30 years ago, where suddenly they had to get on the bandwagon of the web or get left behind,” Weber says.
But there are also key differences. The explosion of the internet in the 1990s bred a feeling that anyone could start the next world-changing website, Weber observes. But with AI, he adds, “there’s more of a concentration of power” due to the vast resources required to train the models and run the technology—even if new open-source models provide individual entrepreneurs a degree of opportunity.
As part of its historical perspective on AI, “Chatbots Decoded” aims to remind visitors that the technology is a human creation, and that it’s up to humans to responsibly guide it. “It’s humans that refine and train all this, and it’s based on human data. It’s not magic; it’s more of a mirror, a reflection of us,” Weber says.
“AI is people, to a surprising extent.”
“Chatbots Decoded: Exploring AI” opened Nov. 20 and is expected to remain open for two years. The Computer History Museum is located 1401 N. Shoreline Blvd., Mountain View, California. Museum hours are Wednesday through Sunday, 10 a.m. to 5 p.m. General admission is $19.95, with discounted options for seniors, students and youth.