The Algorithm for Love

AI is hot right now. Everybody speaks about Artificial Intelligence, either from how is dangerous and will leave millions unemployed, or how will help to save this world, fixing one issue at a time. Truth is, AI gets more involved in our ordinary lives with each passing day, either willingly or unwillingly accepted. So, question about AI being fully incorporated into our lives is more a case of “when will happen” rather than “if it should.”

As AI invades of our lives with the promise of modernity and improving our quality of life, certain areas would be rapidly taken over by machines, as Chat GPT4 has already proven in the text generation space, but other more complex tasks will certainly take longer. But scientists believe that no endeavour will be too big for AI to execute in a near future, as the computing power and algorithms used to predict behaviours get more sophisticated.

So, having that premise in mind, is that I dare to wonder, would be AI able to use an algorithm for love which helps people to find their significant other in the future?

But to answer this question, we first need to understand what an algorithm is and how are used by “AI” to predict behaviours.

I’ve learned to use and design algorithms as part of my Data science studies to find solutions or predict behaviours mostly from the marketing and business decision-making point of view, but I want to clarify that math is universal in that regard, so the principle is the same.

Algorithms are a sequence of rigorous instructions to solve a problem. And the problem in question here is: how we pair two people with unique characteristics with confidence to ensure they will have a happy/lasting relationship? From the mathematical point of view, this is an “optimization” problem, which means the intention is to find the best solution from all available solutions. Here, pairing with different potential matches to identify the highest satisfaction in the relationships.

In the way an algorithm like this could be structured would be first to identify the variables (attributes) to consider in the study, and here is where things can get very complex. Speaking about human beings, these variables or characteristics can be demographic (age, race, gender, ethnicity, religion, income, education level, etc.), psychographic (personality, lifestyle, social status, activities, interest, opinions, attitudes, etc.), behavioural (patterns, habits, how they react to stimulus, etc.), and the list can go on and on. And here we find the first issue faced when coding our Algorithm, the intention is to determine how much these “independent” variables or attributes can “explain” the desired behaviour, which is for two persons to be “happy” in their relationship, but the number of variables we can track would certainly affect the effectiveness of our algorithm. In this case, the more exhaustive the better.

So, this would imply that we would need to build and maintain a humongous database to track each one of these variables to determine if these have a meaningful impact on the outcome. We could argue that according to Moore’s law, computing power “doubles” every 18 months, and with “quantum” computing, this rate will increase exponentially, so again is more a question of “when” than “if it will happen.” Right?

But here is where the second problem arises, just like computers, human beings are getting more complex with time as well.

If we want to reproduce a list of “relevant attributes” for picking a partner back at the time of our grandparents, the list would have been slimmer. Back then, they relied more in a “role” dynamic of men being the “provider” of resources and security and women of “caring” for the house and children, and even many times “love” was out of “marriage” equation. But with our parents, things changed as both genders started to pursuit their freedom, individuality, and craft their own concept of love and relationships. Now, in our times, our personalities have become drastically complex, as our generation leans toward the continuous search for immediate pleasure, conflicting with the growing pursuit of meaning and transcendence.

As our personalities fragment, our puzzle piece becomes more irregular by adding extra knobs and sockets, making it harder for us to find a fitting puzzle piece. Social trends like singleness, voluntary or involuntary celibacy or even the increase in divorces, might have a direct correlation with the increasing complexity in personality hindering human relationships.

Dating APPS like Tinder or Bumble and some people might be disheartened of hearing that regardless of the progress accomplished in the computing and mathematics fields, the construction of an Algorithm for Love powered by AI and Machine Learning, will be only relegated to fiction for the years to come.

However, the study of these models is still of interest in psychology, to help patients to identify patterns which could aid in their relationship therapy.

But what do you think? Should scientists focus on developing an Algorithm for Love? And if so, would you use it?

M. Ch. Landa

Scroll to top
Copy link