Close encounters of the AI kind: The increasingly human-like way people are engaging with language models

Close encounters of the AI kind:
The increasingly human-like way people are engaging with language models

Elon University’s Imagining the Digital Future Center has released a new national survey finding that 52% of American adults now use artificial intelligence large language models (LLMs) like ChatGPT, Gemini, Claude and Copilot. This makes LLMs one of the fastest – if not the fastest – adopted technologies in history.

LLM users say they have human-like encounters with these tools, and some users say the models have personality traits that are both pleasing and dismaying. For instance, among LLM users:

• 65% say they have had spoken conversations with back-and-forth interactions with LLMs, including 34% who regularly do this at least several times a week.

• 49% think the models they use are smarter than they are.

• 40% say the LLM they use most acts like it understands them at least some of the time.

• 32% say it seems to have a sense of humor.

• 25% say it acts like it makes moral judgments about right and wrong.

• 25% say the LLM they primarily use cheers them up.

• 22% say their primary LLM seems to express empathy.

At the same time …

• 50% of LLM users say they have felt lazy when they use LLMs.

• 35% have felt they were cheating.

• 35% have felt frustrated or confused.

• 33% have felt they were becoming too dependent on LLMs for answers.

• 23% say they made a significant mistake or bad decision by relying on information generated by LLMs (23%).

• 21% felt manipulated by the LLM.

This study challenges the idea that LLMs are mainly a work tool. In fact, personal uses dominate the use of LLMs:

• 51% of LLM users say the main purpose for their LLM use is personal, informal learning, compared with 24% who say the main purpose is for work activities.

• 34% use an LLM at least once a day, including 10% who say they use the tools “almost constantly.” Another 18% say they use LLMs several times a week.

• 72% have used ChatGPT, while 50% have used Google’s Gemini, 39% have used Microsoft’s Copilot, 20% have used Meta’s LLaMa, 12% have used xAI’s Grok, 9% have used Anthropic’s Claude and 9% say they have used other LLMs.

• 54% say their use of LLMs has improved their productivity a lot or somewhat.

• 42% say their use of LLMs has improved their creativity a lot or somewhat.

LLM users have a generally worrisome sense about how the models will be socially harmful over the next decade. But they are also relatively confident LLMs will lead to major new medical and scientific breakthroughs and that they will be controlled by human beings and responsive to human needs.

Get the full survey details at the links below.