ChatGPT was launched on November 30, 2022, ushering in what many have called artificial intelligence’s breakout year. Within days of its release, ChatGPT went viral. Screenshots of conversations snowballed across social media, and the use of ChatGPT skyrocketed to an extent that seems to have surprised even its maker, OpenAI. By January, ChatGPT was seeing 13 million unique visitors each day, setting a record for the fastest-growing user base of a consumer application.
Throughout this breakout year, ChatGPT has revealed the power of a good interface and the perils of hype, and it has sown the seeds of a new set of human behaviors. As a researcher who studies technology and human information behavior, I find that ChatGPT’s influence in society comes as much from how people view and use it as the technology itself.
Generative AI systems like ChatGPT are becoming pervasive. Since ChatGPT’s release, some mention of AI has seemed obligatory in presentations, conversations and articles. Today, OpenAI claims 100 million people use ChatGPT every week.
The success of ChatGPT speaks foremost to the power of a good interface. AI has already been part of countless everyday products for well over a decade, from Spotify and Netflix to Facebook and Google Maps. The first version of GPT, the AI model that powers ChatGPT, dates back to 2018. And even OpenAI’s other products, such as DALL-E, did not make the waves that ChatGPT did immediately upon its release. It was the chat-based interface that set off AI’s breakout year.
There is something uniquely beguiling about chat. Humans are endowed with language, and conversation is a primary way people interact with each other and infer intelligence. A chat-based interface is a natural mode for interaction and a way for people to experience the “intelligence” of an AI system. The phenomenal success of ChatGPT shows again that user interfaces drive widespread adoption of technology, from the Macintosh to web browsers and the iPhone. Design makes the difference.
At the same time, one of the technology’s principal strengths—generating convincing language—makes it well suited for producing false or misleading information. ChatGPT and other generative AI systems make it easier for criminals and propagandists to prey on human vulnerabilities. The potential of the technology to boost fraud and misinformation is one of the key rationales for regulating AI.
ChatGPT is not the first technology to be hyped as “the next big thing,” but it is perhaps unique in simultaneously being hyped as an existential risk. Numerous tech titans and even some AI researchers have warned about the risk of superintelligent AI systems emerging and wiping out humanity, though I believe that these fears are far-fetched.
The media environment favors hype, and the current venture funding climate further fuels AI hype in particular. Playing to people’s hopes and fears is a recipe for anxiety with none of the ingredients for wise decision making.
This slowdown should give space for norms in human behavior to form, both in terms of etiquette, as in when and where using ChatGPT is socially acceptable, and effectiveness, like when and where ChatGPT is most useful.
ChatGPT and other generative AI systems will settle into people’s workflows, allowing workers to accomplish some tasks faster and with fewer errors. In the same way that people learned “to google” for information, humans will need to learn new practices for working with generative AI tools.