Since the launch of ChatGPT in November 2022, many people in science, business, and media have become fascinated with artificial intelligence. A cursory look at my own published work during that period shows that I was one of the guilty. In my defense, I believe, as do other obsessives, that large language models are at the forefront of epochal change. Maybe I’m steeped in generative love, but I believe that the advances in artificial intelligence at our disposal will not only change the way we work, but also the structure of businesses and, ultimately, the course of humanity.
Not everyone agrees, and there has been a backlash in recent months. Some experts now believe that artificial intelligence has been overhyped and overhyped. Gary Marcus, the self-proclaimed leading critic of artificial intelligence, recently said of the LL.M. craze, “I wouldn’t be surprised if at some point this all fails. .” Others claim that AI has hit a “trough of disillusionment.”
This week we got some data that doesn’t answer the larger question but provides a snapshot of how the United States (if not the world) views the emergence of artificial intelligence and large language models. The Pew Research Center, which conducted similar surveys during the rise of the internet, social media and mobile devices, released a study on how ChatGPT is used, viewed and trusted. The sample was collected between February 7 and 11 this year.
Some numbers might seem at first glance to suggest that the LLM controversy may be a narrow disagreement that most people don’t care about.one-third of americans never heard of that ChatGPT. Less than a quarter have used it. Oh, and the panic about how artificial intelligence will flood the public square with misinformation about the 2024 elections? So far, only 2% of Americans are using ChatGPT to get information about the presidential election season that has begun.
But more broadly, the survey data suggests we are seeing a powerful technology whose rise is only just beginning. If you accept that the Pew Research Center sample is representative of all Americans, then millions of people are indeed familiar with ChatGPT. One thing in particular stood out: While 17% of respondents said they had used ChatGPT for entertainment, and a similar number said they had tried using it to learn new things, a full 20% of adults said they Already using ChatGPT for work. This is a significant increase from the 12% who responded in the affirmative when asked the same question six months ago – an increase of two-thirds.
When I spoke with Colleen McLean, a Pew research associate who worked on the study, she agreed that this seemed consistent with other huge technological changes. “If you look at our trend charts over time for internet access, smartphones, social media, there’s definitely this uptick in some of those trends,” she said. Some technologies have plateaued, she added. But in the examples she mentioned, plateaus only occurred when too many people joined and few were left behind.
The crazy thing about the sudden jump in ChatGPT business usage from 12% to 20% is that we are only at the beginning of humans working with these models. Tools for taking full advantage of ChatGPT are still in their infancy. This situation is changing rapidly. OpenAI, the creator of ChatGPT, is firing on all cylinders, and AI giants Microsoft and Google are still shifting employees and redesigning every product line to integrate conversational AI. Startups like Sierra that build agents for enterprise customers are enabling custom uses that leverage multiple models. As this process continues, more people will use AI tools. Since the base models are growing exponentially (I hear GPT5 is coming this year?), this will make them even more compelling. This raises the possibility that the quality of almost all work will depend on how well people use the talents of robotic collaborators.
What technologies from the past can help us understand the trajectory of the rocket ship we’re in? Although the upper limit of artificial intelligence is almost unlimited, so it is difficult to find similar products, I recommend using spreadsheets. Dan Bricklin and Bob Frankston invented them in 1978, and a year later the concept was embodied in VisiCalc, which at the time ran only on Apple computers. Spreadsheets have had an astonishingly disruptive impact on the business world. More than just accounting tools, they ushered in an era of business innovation and shook up the flow of information within companies. However, it took several years for spreadsheets to become widely adopted in the business world. The turning point came with new, more powerful products called Lotus 1, 2, and 3, which ran on IBM PCs. Current and near-future startups in the field of artificial intelligence, such as Sierra, hope to become the Lotus of our time, but also hope to become more important and lasting. Spreadsheets are largely limited to the business world. LLM seems to mess up anything.