10 Lessons AI Taught Me To Improve My Mental Health
The Problem with Epiphanies
I remember having epiphanies in my graduate school AI and statistics lectures. The topic that changed my life forever?...Randomness. I was sitting there writing out a pseudorandom number generator algorithm by hand and started thinking about statistical tests for randomness and how we couldn’t replicate “true” randomness and how we’re only hunting for patterns we think exist and then well light particles are random but are they really random or we just think they’re still random and ahhhh….cue an existential crisis for the next week of my life.
Scenarios like this are just a fun peek into how fast my mind could spiral down the rabbit hole and this was waaay before 2020. Let’s extrapolate this out to the number of mental rabbit holes I experience in my career, my relationships, my friends and family, my self-worth, my finances, my outward appearance, and everything in between and you can see how things can get messy very quickly.
After over 10 years in the tech industry coupled with just life in general, throughout 2019 (and of course into 2020), my mental health was suffering - a lot. And I was at a breaking point - physically, mentally, emotionally, and spiritually.
None of the methods I tried before worked long term for me. I would talk it out or compartmentalize, feel better, and then right down the mental rabbit hole again a few days later. I needed something that would stick.
I started to look at solutions that came from within myself.
I was over in Germany giving a lecture on AI technology and while I was crying in my hotel room later that night, I started thinking. AI is something I know, have studied, and think about everyday.
I asked myself...
How can I use what I know to help me overcome my own mental health issues that are uniquely my own?
Thought Experiment: Lessons Learned
Lesson 1: I have to normalize uncertainty.
AI technology forces us to accept the uncertainty of the predictions it makes.
If I reframe my perspective to operate in uncertainty as its normal state, then I can start to conquer the fear of the unknown and expect my decision-making to be imperfect.
Lesson 2: My data will never be perfect.
The data used to train AI technology will never be perfect. Bias of all kinds will always exist in some form, missing data will happen often, and even meaningless data gets collected all the time.
If I reframe my perspective on my experiences to include that my conscious and unconscious bias from my personal life experience is inevitable, that I will never know the full story or motives of everyone I meet, and that I can choose what information I deem meaningful to me in my life, then I don’t have to inflict myself with guilt or ideas of perfectionism on how I respond to life and instead focus on opening my awareness.
Lesson 3: I can use more than one algorithm to answer the same question.
In AI, there are multiple algorithms that can accomplish the same task or answer the same question in different ways. The algorithm used in one solution may be different than an algorithm in another solution, even using the same data.
If I reframe my perspective to understand that different people can solve the same problem in different ways even if they are given the same information, then I can always remind myself that “my” solution is only “a” solution, not the “best” or “only” solution.
Lesson 4: I need to go deep to learn more.
In AI, as deep learning systems learn more about the data, each level of learning transforms its input data into a slightly more abstract and blended representation.
If I dig deeper into certain subjects and look under the hood on topics that on the surface may seem straightforward, finite, or absolute, then I can begin to connect my logic to more abstract representations that can result in blended concepts that allow my experiences to become more powerful and meaningful than what I originally thought or expected.
Lesson 5: Even if my vision is clear, I can still make mistakes.
In AI computer vision technology, even if the computer can “see” physical reality clearly, it can still draw the wrong conclusions or make the wrong assumptions. For this I like to use an example from a colleague’s story when his CV program was accurately detecting wolves 99% of the time until it tried to detect a wolf during summertime instead of winter. Turns out the AI was seeing snow instead of wolves.
If I reframe my perspective to see that even if I have a clear vision for my life as a result of healing my past trauma, I will still make mistakes on the details of my journey to that vision, that vision can be changed, and new traumas will continue to pop up.
Lesson 6: I’m a human, not a robot.
In AI, hopefully self explanatory.
If I reframe my perspective to see that I am not a robot, then I can start to embrace my emotions, vulnerability, and need for rest, setting healthy boundaries on my time, energy, and work that I give to others.
Lesson 7: I have a lot of hidden layers.
In AI, deep learning involves a lot of hidden layers between the input and the output in a neural network.
If I reframe my perspective to understand my interactions with the world are filtered through hidden layers within me that I don’t realize, then I can start to fall in love with my unique individuality that others can’t see, instead of trying so hard to fit in and be upset when others do not see or understand the hidden layers within me.
Lesson 8: I need positive reinforcement for minimalistic and balanced actions.
In AI, reinforcement learning focuses on taking actions that seek the maximum reward. It also focuses on exploring uncharted territory without assumptions while still using the knowledge it has already learned to benefit it
If I prioritize my actions based on impact and explore new areas of interest without abandoning or disregarding the knowledge I have already gained, then I can start to do less with more meaning and apply my prior knowledge to new territories to enrich new areas of interest. This view brings positive reinforcement to taking risks and exploring new things instead of viewing them as a lack of knowledge or something missing from my life.
Lesson 9: I need to learn on my own AND with others to accelerate.
In AI, a single AI use case, API, or application can be powerful, but when AI interacts with the other AI use cases, APIs, and applications, the network accelerates its learning, impact, and intelligence at a quicker rate.
If I reframe my perspective to understand that while I can go far on my own and celebrate and enjoy my individual pursuit of knowledge, connecting with and learning from others that are focused on other bodies of knowledge will help accelerate my learning and understanding to a degree that is impossible to do on my own. This view highlights my connection with others as an acceleration path instead of again viewing it as a lack of something within myself.
Lesson 10: I don’t have to be scared of technological progress.
AI is not the answer to all of our business, technological, or human problems. It is a piece of the puzzle and is only complementary addition to humanity, not a replacement for it.
If I view AI technology as a mirror of the human mind and our collective obsession with trying to uncover the secrets of humanity, its significance and scare factor lessen for me. I can start to see it as a humanities discipline that reflects our pursuit of understanding as opposed to a technology discipline that is hollow and detached from a meaningful human experience.
Upon Reflection
These 10 lessons have been deep and transformative for my personal development. At times, I questioned my direction of studying AI and contributing to the AI and technology industry. I had to grapple more than once with moral decisions of technical strategy on global scales and it took its toll on me.
Reframing my perspective and examining how I could use AI to teach me how to improve my mental health and see the light in my times of darkness has been a personally noble pursuit that I will never take for granted.
If you are reading this, I hope these lessons can be of value to you or encourage you to take some time to see how you can tap into your own knowledge bank and see how it can apply to any mental health issues you may be dealing with.