In the last months, once again AI became a hot topic, after several papers regarding art made by artificial intelligence, now is the time for text written by AI to shine.
Machine learning models were already helping people for a long time, such as grammar checkers, tools to rewrite text to avoid plagiarism, and even tools that do it all - as long as you know how to ask. Tools like ChatGPT aren’t new, they just exploded in popularity recently, and you certainly have heard about many ways to use these artificial intelligence tools, but I want to propose a new way in which we can take advantage of them.
Looking stuff up on the internet can be hard, and sometimes you need a lot of information regarding what you’re looking for, which makes your journey slow. In these cases, using AI tools can be helpful. Unlike search engines like Google, artificial intelligence tools are capable to understand a question and give you an elaborated answer without the need to use special “codes” or filters, you just need to use natural language.
Recently I was looking for how to find circles on images using Python. All the search results on the first page of Google returned an answer containing the package OpenCV, which is one of the main packages in the area of Digital Image Processing, but I wanted something different.
Even though I changed the parameters of my search, the search engines kept insisting on showing me pages with the package opencv, even when I explicitly wrote “without using opencv”, then I decided to test ChatGPT. First I try asking you to write a function to find the circles.
It did write a functioning function, pretty similar to what I had written, but it still was using OpenCV, then I decided to be more specific and ask for other ways to do this, instead of just asking for the code.
I don’t know if all the methods it suggested work or even exist, but these answers help me break a creative block I had, regarding how to solve this issue, and it made it possible for me to get back on making progress on this project. Now I can learn more about scikit-image, which I didn’t even know, and test new packages. It was satisfactory to be able to find new ways to solve the same problem. Using this I can assure you I’m exploring many different paths and I’m not stuck on just what I already know and use.