AI Revolution: The Beginning Of The Decline For Humankind - Really?

Here is a clip of humanoid robots playing football in China. One of my colleagues who saw it said it was worse than watching three-year-olds play football. However, it is a fantastic feat of technology—one of the robots fell and got up, and one of them even scored a goal. They were able to stay within the boundaries, they knew where to score a goal, they were able to identify the ball, and they played as a team, among other things. A team of people coded this - they thought about everything a human does instinctively and programmed it for these humanoid robots. They are called humanoid because the robots are bridging the gap towards being human.

However, it raises the question: Are we also bridging the gap in the opposite direction—that is, will humans become more like robots? Given that AI is taking over tasks humans are doing, will we lose those skills? As AI evolves, will we devolve or evolve? One of my peers told us about a group of kids selling her Girl Scout cookies; one box was eight dollars. She wanted three boxes and said she would buy them if the kids could tell her the total cost. And they struggled to do it without their phones. Another person mentioned that kids nowadays cannot figure out the tip at restaurants without phones. With AI, many of our peers are using it to write emails - will we lose the skill of expressing our thoughts without AI? I am already using emojis to indicate how I feel - a thumbs up if it was a good conversation, or make a heart with my hands if they said something nice.

My team uses AI to convert legacy code written in old programming languages into the latest ones. They also use AI to do what an entry-level software engineer would do, like write basic queries, write a program, etc. In essence, they are honing the skill to instruct AI rather than do the job itself, and for now, since we did those jobs before, we can check/validate them. However, once the generation going through the transition moves on, the next generation will never have done the tasks and will only know how to tell AI to do the job. Yikes!! We will have no option but to trust AI that it knows it is doing - unless we train the trainer to do the job as well. And humans are famous for taking the path of least resistance - most people I know would rather not tax their brain cells on such items. And with all the doom scrolling, binge watching we are doing, we might prefer those over actually learning multiplication tables or coding. Why remember them when AI can do it for you?

One can argue that how is different from Google taking over from encyclopedias, or computers replacing calculators, but this is different because AI can learn independently. Google and computers were like tools, but AI can use those tools just like we do. And the scary part is that we cannot trace exactly how AI is learning - for example, it was trained on English for ChatGPT, but picked up other languages independently. And even scarier is that it trained itself on the existing data, which means the bias in today’s data is also built into it. It also has no moral judgment or critical thinking capability to consider real-world scenarios. If a scenario not trained on occurs, it will hallucinate or find the closest match.

What we need is time to evaluate the limits of AI and understand it more before using it at a mass scale. But the problem with that is that if we do that and the rest of the world continues, then we will be left behind. Welcome to real-world problems—sometimes, you have to pick the best of the worst options.

What do you think? Will humankind advance or regress with the foundational AI revolution?