Artificial Intelligence
What Will Happen If AI Will Be The Next President?
1 reason why it makes sense, 1 reason why it doesn't
This conversation appeared during a typical wine-infused lockdown-inspired Friday-night with my family. After discussing politics and multiple decisions that have been made throughout history, I’ve asked a logical (in my head) question:
If what it takes to be a good country leader is an ability to make good decisions, would you like an AI to do it?
My question was met with a mix of shock and silence.
As I said before, in my head it is a logical assumption. So I made a research. And found out that people are wondering about it since 2007 (at least online).
Here I’d cover 1 reason why theoretically it might work and 1 reason why it might not.
Why it might work?
Decision-making without human emotions
Emotions are an essential part of our humanity. More often than not, we base all our personal life decisions on our emotions. But even if it’s in our nature, is it appropriate to base the important decisions about the nation purely on something like emotions?
The suggestions made by AI are emotion-less. AI doesn’t have personal preferences towards or against certain matters. AI cannot be jealous, greedy, envious, angry. It cannot be pressed, threatened or corrupted. AI is not able to feel impulses and it doesn’t have an ego. That’s why the suggestions that AI will propose are based exclusively on facts from the deep data analysis and logical conclusions.
All of the country’s external and internal political decisions would be made as a result of an analytical process made by the AI. The algorithm can analyze the tons of historical data in mere minutes to predict the possible and most certain consequences of different political decisions.
AI will analyse tons of data of the previous mistakes throughout the country’s and the world’s history. It can analyse the pattern in people’s behaviour as a response to various new legislations. It can even analyse the psychological and neuroscientific aspect of our behaviour and predict potential public’s responses based on that.
The AI can be programmed in a way that will only consider solutions and decisions that will protect the nation and the whole of humanity. If the initial AI algorithm is programmed in a way to ‘keep the peace’ within and outside the country, this default setting will determine the potentially risky actions and will remove them from the list of suggestions.
It seems pretty straightforward. However, practice is always different from theory.
Why it might not work?
AI cannot explain its’ decisions.
The data is fed to the AI ➡ AI analyses the data ➡ AI makes the decision
The algorithm seems straightforward. But here is a hidden problem. At the current stage of AI development, the algorithm cannot explain its decisions. It cannot show us, humans, why did it come up with this particular decision. We don’t see the decision process. We don’t know why exactly the machine has made this choice.
➡ Why is it a problem?
The AI algorithm can propose a certain decision in a certain situation. It will identify patterns and base its’ decisions on these patterns. The problem is that some of these patterns might make sense only to machines and not to humans. Therefore to collaborate effectively in this area, AI has to be more advanced to be able to show the “flow of thought” and justify its verdict.
Let’s imagine a situation: the government ‘asks’ AI to analyse all of the information there is on the subject of roads repairing and answer the question ‘Should we repair roads in city X?’. Let’s say, the AI answer will be ‘No’. There will be no supporting data regarding ‘Why No?’. Now imagine, the question is harder or more controversial. AI proposes a ‘Yes or No’ solution without explaining Why Yes or Why No.
What is the most possible scenario right now?
The most possible solution is a closer collaboration with AI. AI can provide better guidance on various matters, it can help the governing parties to evaluate their decisions from different perspectives and it can be used to analyse the big amounts of data.
AI can identify the potential flaws and prejudice of every decision while leaving the final choice to the human.
