The first time most people could communicate with a computer that responded like a real person was in 2022 when OpenAI publicly launched ChatGPT. Just two months later, the app set the record for the fastest-growing user base with 100 million monthly active users. 

The service was prone to mistakes and hallucinations, but the conversations felt more or less real. Since then, advances in artificial intelligence (AI) as a technology have been stunning. AIs have passed the Turing test, a standard that determines whether a machine thinks by whether a human believes they are communicating with another human being. AIs score better now than humans on the SAT, GRE and LSAT.

Beyond these seemingly shocking advances, most of the AIs all around us remain out of sight. AIs manage energy distribution across the country. On dating sites, they help us pick potential partners. In the world of finance, about 70% of stock trading volume is initiated through algorithmic trading, another form of AI.

“I would say that AI is part of the human system,” said Martin Hilbert a professor of communication in the College of Letters and Science at UC Davis.

Hilbert is among the many social scientists across Letters and Science who study AI to answer questions about people and society and to develop technologies that may drive our future. This work touches nearly every part of daily life, from health to education to the way we communicate with each other. Their work shows that at this point in history, AI and human society are inextricably intertwined.

“It’s the end of our species if you take seriously what it’s about,” said Hilbert. “It doesn’t mean we’re going to die, but it means there is an artificial element, an offspring of ours, that also possesses the defining aspect of our species. We’re no longer the only being in our universe that has this knowledge business down.”

AI driving the frontier of social evolution

The 2013 film Her tells the story of a man who falls in love with the chatbot embedded in his computer’s operating system. Throughout the film, the voice who speaks to him during nearly every waking moment — like the chatbots that seem imminent — feels like it comes from flesh and blood. 

Except AI is not a person. So what exactly is it?

AI is a type of computer program that works like a very long series of probabilistic if-then statements that rely on data to find the right response. When we interact with an AI, whether it’s on ChatGPT or a chatbot that helps us book a flight online, its responses are based on a database of similar interactions that help it predict what to say. 

This technology that could help to solve some of society’s problems is actually reorganizing society itself. In a 2023 paper on SSRN, Hilbert named this kind of technology-driven reorganization “algorithmification.”

“Algorithmification does to knowledge what digitalization does to information,” said Hilbert. 

Digitization created widespread access and use of information that was once contained exclusively in books, individual databases and human minds. Hilbert argues that algorithmification expands the reach, access and potential of knowledge and information processing that was only done by human minds.

Hilbert explained that this shift is only the most recent iteration of our social evolution, whic has been the primary way humans have evolved for the past 10,000 years. That change has been driven by technology, and it has affected every living person in human societies.

“You could get rid of your cellphone and you could never touch money, and you could go out into the desert and maybe you could survive for a few weeks or a month,” said Hilbert. “You would still be a human but you would be like a human thousands of years ago. You would not be evolving with the rest of us in society.”

AI and the future of business

When it comes to the economy, there’s a lot of fear about AI costing people jobs and even making some jobs obsolete. Hilbert says some of this is true. The key, he said, is to understand how to use AI technologies for your benefit. 

While entire job categories are unlikely to be replaced, Hilbert also said that workers who don’t use AI will likely be replaced by workers who do. However, AI will also make it possible to significantly accelerate contributions to education, medicine and so many other areas by its ability to produce language and software code.

“While we have put much effort into writing code for marketing and security, we haven’t really started yet writing code for education and healthcare or for arts and entertainment to the same degree,” said Hilbert. “We still need thousands of teachers around the world. Doctors, the same thing. If there are certain tasks that can be alleviated with algorithms, that would be amazing.”

The promise of AI has captured the imagination of venture capital firms investing in the future. UC Davis alumnus Vince Lane is a venture capitalist investor with a focus on AI and other “frontier technologies” like software-as-a-service (SAAS) and cybersecurity. An entrepreneur with over 25 years of startup and corporate experience, Lane is also a member of the Letters and Science Dean’s Advisory Council.

“If you don't utilize artificial intelligence then you aren't even being looked upon as a viable company,” said Lane. 

Lane saw this kind of disruption in the 1990s with the earliest widespread access to the world wide web, but said he was too young and early in his career to be a major part of it. Today, he is at the center of AI innovation, helping to steer its course through strategic investments and advice.

Lane said that AI is going to break through the limits of what we can accomplish on our own. This includes speeding up technological and bio-medical advancements. It also includes streamlining functionality for logistics and software development. He said that these kinds of innovations are going to make everyone more efficient and happier in their day-to-day lives.

“With artificial intelligence, I will say that even though I'm so ingrained the whole community is moving very, very fast. I think it's one of the biggest disruptors of our century,” said Lane. 

Interacting with AI in daily life

Our lives are increasingly going to involve interactions with AI, for better or for worse. These interactions in particular are a focus of research about how we can be sure those interactions help us as individuals and as a society.

“We are going to increasingly interact with AI models in the future,” said Jingwen Zhang, an associate professor of communication. “We have to really think about the ideal scenario of how humans interact with them and how AI models are designed to benefit humans.” 

Zhang studies how AI can help improve public health, and she has been experimenting with a chatbot that can motivate us to act. She and her research team designed an AI chatbot that crafts personalized, persuasive messages that help us make sure to get in our steps every day. The chatbot can be downloaded via app on any device. 

In Zhang’s first field experiment with the technology, the AI chatbot delivered a single 10-minute conversation to the study’s participants. That single conversation increased people’s intention to exercise by 37%. According to Zhang, increasing the duration and dynamic nature of the interactions could have a much stronger effect.

Kenji Sagae, professor and chair in the Department of Linguistics, studies computational linguistics, and his focus is on how large language models (LLMs) impact society from a language perspective. A language isn’t only a vocabulary of words and structures of grammar to make meaning out of them. A language conveys culture.

In one project, Sagae is studying the cultural values contained in LLMs in terms of moral values. In one study, he and his research team have been asking participants to fill out questionnaires that help to define their values, which differ across cultures. 

Sagae is measuring variations in cultural morals with the Moral Foundations Questionnaire that can score an individual’s values in terms of harm/care, fairness/reciprocity, in-group/loyalty, authority/respect and purity/sanctity.

One of his findings is that LLMs are most likely to contain the moral values, or biases of the training data. For example, if the LLM is trained on web pages from the U.S., the embedded values in how it communicates with people from any part of the world will represent values from the U.S.

The way we communicate with AIs might also affect how we communicate with each other, and Sagae said that this isn’t by design. It’s a side effect. This is critically important, he said, because we are in a moment where more and more of our everyday interactions are with computers and not individual human beings. This could have its biggest impact on children, who are still learning how to communicate through language.

“People are not looking seriously yet at what it means for kids to communicate with machines,” said Sagae.

Children might grow up learning how to use language outside of a community context, which has always been the way we learn languages. Instead, they could learn more and more  how to use language from an entity that has no emotional connection to the world. 

“There is certainly potential for us to benefit greatly from the technology,” said Sagae, “but there is still a lot of work needed to get us to the understanding that will unlock that potential without unintended harm.”