What Should Users Make of New Google AI Tool? We Asked a UNH Expert

Google recently announced the release of Personal Intelligence, aimed at enhancing the company’s Gemini AI assistant by allowing Gemini, with the user’s permission, to access the user’s Gmail, Google Photos, YouTube, and search to better anticipate how to assist the person. So, how does Gemini’s newly added personal intelligence compare with our human personal intelligence? What does this technology mean for users, and what are the potential advantages and pitfalls of using it?

We spoke to John (Jack) Mayer, psychology professor and recipient of UNH's 2025 Distinguished Professor Award, to learn more. Mayer is the author of the book, “Personal Intelligence: The Power of Personality and How it Shapes our Lives,” and his most recent work has been concerned with how human personality is changing as people use artificial intelligence.

What is your background working in the area of personal intelligence?

I published a theory in a 2008 article that defined personal intelligence as the ability to understand personality in oneself and others, and laid the foundation, more specifically, for its measurement. In 2012, colleagues Abigail Panter, David Caruso, and I published our first test of personal intelligence. I went on to write a book about what we had learned about it to that point.

In 2014, I was invited to present at “Talks at Google” on the theory, where I described what the intelligence was and how I believed people used the intelligence to make key decisions in their lives. 

A number of former UNH lab members became involved in the work shortly thereafter. Jayne Allen found that people high in PI, relative to those lower, engage in more sophisticated learning about others’ behavior. Kateryna Sylaska showed that college students high in personal intelligence choose majors better aligned with their personalities than those lower in it. Another colleague, William Skimmyhorn, and I found high-PI students do better in courses having to do with human problems — leadership, philosophy, and literature. In continued work with Caruso and Panter, we have shown that high PI students also are more likely to navigate relationships in supportive, respectful ways — and yet to be critical of their friends when needed. And employees high in PI are more aligned with their work colleagues and organization’s goals.

What is Google’s Personal Intelligence purporting to do?

Google says that AIs with personal intelligence collect information about you to learn about who you are and use it to anticipate your needs and goals. By doing so, its developers hope it will make personalized guesses about how to support you in what you want to do.

Can it actually achieve that objective? Are there any potential limitations to the technology in your eyes?

Over the past few years, large language models like Gemini have developed the ability to correctly answer questions like those on our measures of personal intelligence. And, for some time, psychological research has also confirmed that AI systems can extract personality-relevant information about individuals from their online behavior. Just as people evaluate one another’s personality in everyday life.

But there are also key differences between human and AI personal intelligence: We humans have access to our individual feelings and inner thoughts — although to be sure, an AI system such as Gemini has access to millions of human discussions of inner feelings.  But a human observer still has better access (for now) to real life person-to-person interactions, and people’s offline behavior. 

Gemini, on the other hand, has a clear memory (if one allows it access) of one’s email searches, pictures, and viewing habits on YouTube. Because the information that a person draws on, and that Gemini draws on, are different, the person and the AI are likely to come to somewhat different views of the person. Exactly which is better is an interesting question.  

Does providing access to that kind of personal information give you pause for any reason?

This capacity of AI to exhibit personal intelligence raises at least two concerns. First, there are issues of personal privacy and the risk one runs allowing the app to absorb all that online behavior. Second, keeping mentally alert involves using one’s problem-solving skills in an ongoing basis. When you stop paying attention, you may lose practice in the area, and your skills might decline. In a blog post, Google’s Josh Woodward (VP of Google Labs, Gemini, and AI studio) used the app to analyze his family’s past trips and emails to come up with suggestions for a vacation over spring break. It suggested traveling by train, a suggestion Woodward found novel and appealing. (I also have asked AI about vacation ideas.)

If we repeatedly ask AI to think for us, however, we run the risk of losing skills in certain areas — not just vacation planning, but any area we delegate, from reading, to writing, to personal planning. We also cut down a bit on our personal interactions – after I asked an AI about vacation ideas, I followed up with the same question to a fellow gym member, who I both enjoyed talking with and got some more useful advice about a vacation spot in New Hampshire.

What does all of this mean for users of this technology?

In the case of AI with personal intelligence, the boundary between one’s personal preferences and those of the AI could potentially blur the more we use it, but our own gut sense will still matter at the human level. If we increase our use of AI, we may also want to increase our self-understanding to stay in contact with our preferences and desires independent of something else’s neural nets. The approach I would recommend for now is to be aware of protecting your privacy, and if you wish, try it out, but proceed with caution — and remember, if you already possess reasonably good personal intelligence, which most people do — there is some important information about who you are that you are likely to know better than the AI.