Is AI hiding the knowledge?

June 2, 2024 - 6 minutes

Is this a real problem?

Everybody is using AI systems to know more and solve problems. Chat-GPT has beaten the barrier between the skepticism against this technology and now millions of people have been using it every day. As a developer, sometimes I use it to solve some problems in a very short time because it has been improved and now it is very simple to use. Another example is GitHub Copilot. It is a really incredible tool for coding. Once you try it you wouldn’t ever come back and developers definitely will use as much more in the next years. Everything looks simple and efficient, but there is a hidden problem that comes up in my mind. Is it possible that using all these AI tools will contribute to hide the knowledge?

In my experience as a developer, I notice that it is easier to solve a problem with an AI tool rather than search online. In the tech field, everybody knows forums like Stack Overflow or GitHub Community - specifically for developers - or other platforms like Reddit or Quora - to be more generic - and they are crowded with people who like helping other people, and this is amazing. But with the advent of AI tools, is it possible that people use less forum or divulgation website (like Wikipedia or Medium)?

I know that it would be necessary to go in depth about the issue, but unfortunately I have very little data available. I want to show you anyway how the web traffic is moving in the last 12 years on the main platform of community divulgation.

GitHub: Medium: Reddit: StackOverflow: Wikipedia: OpenAI:

From the plots (source: semrush.com) we can see there is a trend, where most websites reach a peak around 2020 and there is a decrease in web traffic. The websites that are growing are Reddit, OpenAI and GitHub. However, it is very clear that OpenAI is increasing very fast, faster than other platforms.

It is not possible to attribute those movements to OpenAI, of course. I think that the peak of 2020 is due to the COVID pandemic where most of the people were at home and they were using the computer to break time. However, it is possible to see a correlation between the increasing of OpenAI and the decreasing of the others, albeit light.

I did this little research to describe a phenomenon that is happening. I think that the utilization of AI tools is contributing to decreasing the community knowledge. Take a second and try to think about what is going on when you are using Chat-GPT. Let’s take this scenario as an example.

You are writing the code of your new application. You are getting mad with a problem and after you try with several attempts you decide to find a solution online. You are tired and you want a solution as soon as possible, so you open your dear and loyal friend Chat-GPT. You describe it your problem and it provides you a good solution. You come back on VS Code and you try to implement it. It works! You are happy and you continue working. But nobody will know how you solved it.

Stop using AI?

Is there something wrong? Apparently, no. But if we think that some years ago, we went on an online forum and we described the problem to other people, life has completely changed. The strength of the online forums was that if I have a problem and someone helps me to solve it, then, it is possible that someone else will solve their problem thanks to the contributions. And maybe that person will understand how you solved it because they will read the discussion and they can learn something new.

I’m not saying we shouldn’t use Chat-GPT, I use it too. But I want to describe better a future scenario. Imagine that in 20 years, developers will be using completely different technologies. Every time we encounter a problem, we turn to Chat-GPT for a solution.

The issue is that while the AI knows how to solve the problem and understands the motivation behind it, people might not. We are training AI to solve problems and we are providing a huge amount of data to AI in exchange for solutions. We are increasing the knowledge of AI rather than increasing our own knowledge. In a way, AI is making us dependent on it. One day in a dystopian scenario, it might happen that AI stops to answer us and leaves us without means to solve the problems. AI could potentially take control and stop to assist us.

Achieving the result: easier or worse?

The AI is making doing everything terribly easy. It provides quick answers to hard questions that once required a lot of study. Have a problem with your code? Ask AI how to solve it. Need to write a script that does this. Let’s ask Chat-GPT to provide it. You need to prepare a presentation but you don’t know how to start. Use AI! Need to create a simple logo for your project. Let’s generate it!

These are just a few examples of how to solve simple problems with less effort and time, but also less study. I don’t want to demonize AI, but if on one hand it allows to save time on boring tasks, on the other AI represents the result of a society that is trying to get faster in every aspect. The important thing will always be more achieving the result instead of how to achieve the result.

In the past, every activity required significant time and effort. Now, in our frenetic world, everything is quick. The deliberate practice cited in “So Good They Can’t Ignore You” by Cal Newport, is getting so hard to play out. Let’s think about old jobs, like carpenter, glassmaker or shoemaker. In those activities you need to work and learn for a long time. You continue to improve every day, and only when you work on it for years, you can consider to be a pro in that field. One activity, for years.

I fear that this will be almost impossible in the coming years. The technology that has been developing is making everything so prefabricated and standardized. While this makes things easier and quicker for everyone, this leaves no room for creativity.

I believe there is no definitive solution to this situation. However, one good approach is to take our time, and resist the urge to rush everything. In this frenetic world, it’s essential to remember that the process of learning is as important as the results we achieve.