Connection Found Between Personalities and Eye Movements Using AI

We know that a person’s eyes can be expressive, but they’re not only showing how we’re feeling, but our overall personality as well according to machine learning research. It was discovered that this new technology can detect four of the “Big Five” basic personality traits when looking into someone’s eyes.

The Research

Artificial intelligence was used to track and analyze the eye movements of 42 students. It was able to detect the personality traits of agreeableness, conscientiousness, extroversion, and neuroticism. What it can’t pick up is openness.

It’s believed that this could improve how people and machines interact. If the machines can learn about our personality traits, it may be able to learn how we would react to something and then do what we need it to do based on that.

“People are always looking for improved, personalized services,” said Tobias Loetscher, University of South Australia senior lecturer. He was the lead researcher for the global team that included his university, the University of Stuttgart, Flinders University, an the Max Planck Institute for Informatics in Germany.


“However, today’s robots and computers are not socially aware, so they cannot adapt to non-verbal cues. The research provides opportunities to develop robots and computers so that they can become more natural and better at interpreting human social signals.”

On the downside, this also implies that this technology could be implemented when people are unaware, causing privacy concerns.

Look what’s happening right now with facial recognition and how it’s now being used by the police to track us down. Imagine what could be done if they started tracking our eyes to learn how we were going to react.

What Your Eyes Say About You

The personality traits have been connected to the following movements:

  • Look around more – curious
  • Stare at abstract images for longer periods of time – open-minded
  • Blink faster – neurotic
  • Moving eyes from side to side more – open to new experiences
  • Greater pupil size changes – more conscientious
  • Spending less time looking at things that are negatively emotional – optimism


These, of course, aren’t hard and fast rules. If you are noticing yourself blinking quickly right now, don’t get worried thinking it must mean you’re neurotic. These are just tendencies they have found.

Being Cautious

As mentioned above, there is a downside. As great as all this technology is when it comes to helping us personally, there is always a chance for it to be used against us.

It doesn’t have to be for nefarious reasons. It might be very well-intentioned, but not everyone wants the intrusion as we have found here.

Again, it’s really interesting that they can determine such things, but do we really want them using artificial intelligence on us to determine our personality and how we’re going to react to a given situation?

Let us know what you think about this research in a comment below. I’ve seen you staring at the screenshots here. You must be open-minded; just don’t start blinking too quickly.

Laura Tucker Laura Tucker

Laura has spent nearly 20 years writing news, reviews, and op-eds, with more than 10 of those years as an editor as well. She has exclusively used Apple products for the past three decades. In addition to writing and editing at MTE, she also runs the site's sponsored review program.


  1. “it’s really interesting that they can determine such things”
    Can they actually? Or are those just one person’s or one group’s agreed upon interpretations? I can provide totally different but still valid interpretations for the six eye movements listed in the article. The problems is that human actions and behavior are open to interpretation and the interpretation depends on the point6 of view of the interpreter. For example “Moving eyes from side to side more” can also be interpreted as “shifty eyed” or “apprehensive”. Or, a child that stands up for himself against others is self-assured. But if the child stands up himself against you, he is arrogant.

    1. The connected Mashable article clarifies this,
      “Although these are fun (and freaky) things to think about, keep in mind that eye movement patterns aren’t determinants for these character traits. These are simply tendencies that the researchers found had a correlation.”

      Also, this was a very small sample size. The devil is in the details. Like how many observations and what was the duration of the observations? Also was there a gender, ethnicity and age breakdown? What affects, if any does wearing contacts have? So this is just the beginning of this research. I’m sure that eye movements and facial expression will be mapped to more cognitive states. So someday (soon) HAL will know what your feeling based on expression, eye movement, and vocal tone.

      1. “So someday (soon) HAL will know what your feeling based on expression, eye movement, and vocal tone.”
        I very much doubt it.

        All this is based on the dubious assumption that human behavior can be quantified precisely, like “2+2=4” or the Periodic Table. Research into human behavior is an art, not a science. The results of experiments are not reproducible. When a wire moves through a magnetic field, the movement generates electricity, EVERY time you repeat the experiment. if you put a human in the same situation 10 times, you may get 10 different reactions.

        1. First of all, sarcasm… If you read the first part most of the research done on people and emotional states is using correlations. So if you 1000 people and map their facial movements (eye movements) or whatever, you can correlate an average response and say something like “…when xzy muscles are engaged people tend to show abc emotion…” That data is fed back into the computer and now it has an algorithm to apply to facial recognition or eye movements. It’s not perfect and for any individual, it could be wrong but it works in the aggregate and enough to be useful.

          1. “First of all, sarcasm… ”

            “So if you 1000 people and map their facial movements etc. etc. etc.”
            If you repeat that with another 100 people, you MAY or you MAY NOT get the same responses. The only way the test is reproducible is if you average the responses out, form them into a trend and apply some fudge factors to de-emphasise the anomalous data points.

            “That data is fed back into the computer and now it has an algorithm”
            That is a self-reinforcing, self-validating feedback loop. IOW, a runaway chain reaction.

            “it could be wrong but it works in the aggregate and enough to be useful.”
            Thank you for proving my point that studying human behavior is an inexact art that depends on approximations, interpolations and interpretations, rather than a hard science. “It could be wrong but it’s close enough for government work.” It’s a good thing that engineers don’t follow that principle when building things. We would be in deep doo-doo.

Comments are closed.