AI in Psychology: Understanding Human Behavior from a Student of Saint Peter’s Preparatory School

Today, I will be discussing the applications of AI in Psychology, and by the end of this blog, you will be able understand how AI makes predictions, how AI is important to the sciences, and how AI is used to analyze behavior.

Image courtesy of Pexels

Predictions are something AI is known for; taking in parameters and outputting a prediction is one of its most frequent uses. Many of you have likely imagined how AI predicts stock prices, evaluates what team would win a sports game, or even appraises paintings! Predictions are an essential part of AI, and here are the basics about how the magic happens:

AI Fundamentals

ALGORITHM TRAINING

AI can be trained by either machine learning and deep learning. The difference between these and traditional code is that they are able to adapt to new data, as you might have guessed. AI is usually trained with statistical models, that use the numbers that they take in to come to a conclusion; the intricacies of this model is where this start to become complicated. There are two common models, linear and logistic regression. The former is used for values, while the latter is primarily used for categorization. However, there are a plethora of models that AI can use, and the more that you know, the better you can train your AI!

PREDICTIONS AND MODELING

AIs utilize neural networks, systems that make predictions by taking in data, make a prediction based on their models, compare their prediction to the actual result, and adjust their model accordingly. These neural networks contain layers through which they run their data. It’s almost like a black box, it’s nearly impossible to know what’s going on within the model. Through the process of trial and error, AI can become incredibly effective at prediction. To train AI, you input a test data set with known outputs, so your AI can check if it’s correct. There is a simple analogy for how AI is trained: imagine you are playing basketball, and you miss the net. After that, you will see where you went wrong; maybe you threw the ball too war to the left, and it missed the backboard entirely. Based on this observation, and knowledge of where you actually wanted to hit, you will adjust, and this will keep going until you are able to come as close to perfection as you can be. AI training is something that many companies try their best to get; some captchas actually use your inputs to train AI! Predictions are just the beginning of how to understand human behavior, but they are indispensable.

AI Applications in the Sciences

Psychology is a science, and AI is associated with technology. Though some parallels can be drawn between those two subject areas, there are some critical uses of AI that have revolutionized science. In biology, AI can analyze protein structures and even fold proteins for research; one such project is trying to discover the entire human genome. In chemistry, AI can predict the properties and structure of molecules, and even what reactants could have made up that molecule. Even in physics, AI can analyze trajectories and create accurate mathematical models. With AI, we can perform research to extents never though possible. If we want to perform an experiment to perfection, and find out the perfect conditions for something, we can calculate that with AI, and it will get better and better as it is trained. The possibilities are endless, and AI has enabled a whole new era of science.

Analyzing and Understanding Human Behavior with AI

Finally, I’ve saved the best for last, and the big question you’ve all been waiting for: how do we analyze human behavior, and learn to better understand ourselves with AI in psychology? Psychology, like other sciences, has one salient requirement: data. Without data, you can’t do anything, and the same goes for AI. Once you have that data, however, you can manipulate it as much as you want to make predictions. AI is actually being used in psychiatry, to treat those with psychological afflictions. According to psychologytoday.com, training AI with the massive datasets that are available can “identify factors that render people more vulnerable to mental illness, improve the accuracy of diagnoses, and assess which treatments are effective and for whom.” AI can help with these diagnoses, as well as to predict the behavior of people. As you may remember, I did a project on AI in criminal justice with Inspirit, and that ties in with psychological AI pretty substantially. Though it doesn’t use psychological traits, it has the basic foundation there. Using various variables, though I excluded ones such as race, gender, and age to prevent unfair bias, AI is able to predict the behavior of criminals, and predict whether or not they would commit another offense. Even with nothing but information about the person’s history and life, AI can draw conclusions that were about 80% accurate. When you add in the experimental nature of psychology, however, you have a force to be reckoned with.

Conclusions

All together, we’ve been able to take AI apart and see how it functions, with many layers and adjustments that fine-tune its calculations. We’ve seen the revolutionary impact of AI in the sciences, the likes of which are completely unprecedented. Finally, we saw how AI is being used in the real world of psychology. Regarding the last point, this is something you can try yourself. Though I didn’t delve into the specifics of the coding, there are numerous ways to code neural networks, and with the right experimentation, you can make a psychoanalytic model of your very own! Though AI cannot drive psychology by itself, as the human mind is far more complex than any of us will ever know, it’s making a splash in the scene, the results which cannot even be predicted by machine learning.

Sources:

https://realpython.com/python-ai-neural-network/#artificial-intelligence-overview

https://www.psychologytoday.com/us/basics/artificial-intelligence

https://www.frontiersin.org/articles/10.3389/fpsyg.2019.00263/full#h6"