AI Algorithms and Life Lessons
If we closely observe AI algorithms, we can draw analogies between the workings of AI algorithms and life lessons — and human nature.
1 // Algorithms don’t rely on “first impressions” or data that is initially fed. If they did, then algorithms would be inaccurate and biased. On the other hand, AI algorithms get updated frequently.
We give too much importance to “first impression”, and use those distorted and inaccurate stories to mislead ourselves until we encounter a huge failure and we have to wake up to the reality and truth.
Key Take-away: Question and update your “first impression” and old beliefs about people, places, experiences, and the world.
2 // Algorithms don’t have “ego”. They learn continuously — yet fiercely — from errors to perform better.
Often, we find it painful to admit we are wrong, and sometimes end up blaming circumstances, find excuses for ourselves, justify our mistakes, and as a result don’t learn.
Key Take-away: Face your mistakes and take accountability for it — even if it is painful. Accept them as gifts and learn from them to create a better version of you.
3 // Algorithms change moderately at a sustained speed. It doesn’t move to fast, otherwise it misses the optimum result. It doesn’t move too slow, otherwise it gets trapped, never updates and fails to learn — given the fast paced changes in its environment.
However, we sometimes go to the extremity of negative conclusions when a thing goes wrong and end up doubting ourselves for our decisions, skills and capabilities. Or we repeat the same work even if it is not working, hoping for it to work.
Key Take-away: Make moderate changes and improvements on a continuous basis. Don’t be too sensitive to failure and doubt yourself, but improve through iterations — however small they are.
4 // Algorithms learn through “interactions and feedback”. They learn to solve a task by “trial and error”.
Sometimes, we prefer to stay within our comfort zone and avoid any interactions, or refrain from providing or taking feedback in a positive spirit. We often circumvent undertaking a trial and error approach only to either end up not solving the problem, or using an existing solution that might not be the most ideal.
Key Take-away: Embrace a collaborative environment and create a diverse team of experts where you learn from one another while providing and taking constructive criticisms and feedback. Get experimental, creative and innovative in implementing solutions through a calculated “trial and error” approach.
5 // Algorithms face the problem of being “over-fit”. But when multiple algorithms work together, the risk is significantly reduced.
We sometimes focus too much on perfection and details which unfortunately might not be worth it, especially for practical purposes. Or, we sometimes choose to own up the task and try to solve it all by ourselves as we don’t trust the work of the other person.
Key Take-away: Collaborative work, taking inputs and ideas from others and leveraging on people’s experiences and advice helps in delivering better results and making better decisions.
While these 5 points talk about the things we should learn from AI algorithms, there are 3 things about them that we shouldn’t be.
1 // Algorithms perform badly — without the right data being fed to it — when dirty data is fed to it
Key Take-away: Use your wisdom, cognition and intuition to go for the experiences, to choose paths, and to make decisions you want to learn from. Don’t let someone else determine your life or make decisions.
2 // Algorithms rely on data they are shown and cannot differentiate between truths and lies.
Key Take-away: Sharpen your acumen and emotional intelligence to identify inconsistencies in attitudes and behaviours, and accordingly act.
3 // Algorithms can work in parallel and in a way can efficiently multi-task due to automation
Key Take-away: We aren’t machines, and therefore should focus less on multi-tasking, and focus more on achieving realistic goals. And focus less on basing our actions on the assumptions of what will happen in the foreseeable future, and focus more on what needs to be done now.
Lastly, a couple of points to take note about algorithms that is influenced through external interventions:
1 // Algorithms can be tuned by us — after careful observation — to provide optimal results.
Key Take-away: Although we are not race cars to tune our brakes and accelerators, we can definitely make the physical and mental shifts on wellness to make sure we can optimally perform and stay healthy.
2. Algorithms learn at their pace. And when multiple algorithms are run in fixed formats, each algorithm picks formats that works well and can then be merged.
Key Take-away: Stay out of the way and allow the learning process to occur without making big changes to the strategy too quickly, while observing competition.
In the background of all this positivity, it is important to inspect the input of potentially biased data to create biased outputs. The data we feed into the AI algorithm is a reflection of our own biases and our personal outlook of human societies. Algorithms are not perfect, just like humans.
Just an idealistic thought — if humans can learn from algorithms, then can algorithms “learn” from humans and one day submit to an appropriate legal framework?
Let me know your learnings and thoughts on this!
