This article was written by Manon, a Python developer at our Toulouse office. Through her experience, she shares her point of view on the impact of AI on the developer's job, and how Artificial Intelligence can be used on a daily basis without replacing human expertise.
"In the age of AI, as a Developer, I've often had the question, "Aren't you afraid that AI will replace your job?". After artists/illustrators, are our professions in danger?", Manon.
Introducing AI and its impact on our daily lives
Artificial Intelligence has revolutionized our daily lives in just a few short years. It has to be said that ChatGPT (the one that really democratized AI) had it all: free and easy to use. In fact, using it couldn't be simpler: just ask a question, and it'll answer you as quickly as a human being would. But don't be fooled by appearances: behind its simplicity lie millions of data points and algorithms for adapting, understanding context and answering questions with relevance.
AI : an ally for Developers and DevOps?
If you're a developer, you'll know that we do a lot of research on the Internet to find the solution to our problems (error logs, for example): we use Google and Stack Overflow, which are very useful sources in our day-to-day work. The AI model lends itself very well to replacing the aforementioned tools. For every problem, there are AI answers. What's more, other AIs such as Copilot offer code review and even code generation.
My experience with AI in my job as a developer
From mistrust to enthusiasm
Unfortunately, my first approach to AI was rather disastrous. All too often, I saw AI misused or over-hyped as being able to solve all our problems. For many months, I was stuck with this very negative image. What's more, I think we developers are somewhat ashamed to talk about the use of AI, for fear of it devaluing the work behind it. I think this subject is still too taboo.
Then, new missions, new contexts, and the first concrete case of AI use. And what a surprise! What a wonderful tool! By using AIs, my research and work time was reduced by a factor of three! A real time-saver, I must say.
The concrete benefits of AI in my work
One question, many answers! Instead of spending hours searching for a bug on Google, I found my answers in five minutes.
For code, this is a very powerful tool! Blocks of code several lines long can be queried, and functions can be simplified very quickly. The AI makes it possible to quickly create frames of code to hang on to. What I also appreciate is the help it gives in naming functions and variables. What's more, it's extremely helpful with documentation and translation. Another useful function is the code sequence suggestion: just as when we start writing an e-mail, the AI suggests logical code sequences.
The benefits of AI : time savings and efficiency
In development, it can really offer considerable advantages, particularly in terms of saving time. Thanks to AI, I've been able to significantly reduce the amount of time I spend not only searching for information, but also coding, generating functions and creating documentation. It simplifies tasks that used to be long and tedious.
The limits of AI: when the human element remains indispensable
A few points to note, however: AI needs to be "well guided" to provide the most relevant answers. If we don't guide it well enough, it will give us generic answers that don't correspond to the context. What's more, I quickly realized that, on certain points, a tutorial or an article will be more relevant than the answers provided by an AI, which has the annoying tendency to forget information that is essential to solving an Ops/Software problem.
As far as I'm concerned, I'm glad I was able to learn to code without these tools, because the errors they can bring to a junior are numerous. Indeed, I found the code she proposed to me sometimes not optimized enough, incomplete, even buggy and wrong! (Yes, yes, it's possible.)
One point to bear in mind is that, as experienced developers, we can quickly analyze the source of errors in the code, taking into account the context where AI has weaknesses. This certainly saves time, but we still have to go back to the drawing board to get quality work.
Protecting of sensitive data in the face of AI
Another point I'd like to address is the protection of sensitive data. As a consultant, it's essential for our customers to protect their data. Every day, we handle information that is the property of the company we are working for. Unfortunately, due to a lack of knowledge or carelessness on the part of some people, too much information is copied and pasted to third-party AIs. I've attended far too many meetings where customers have complained about their staff's misuse of AI.
Solutions
Some companies have chosen to prohibit their employees from using AI within their teams, preferring the preservation of their data to speed. This choice is understandable, as data leakage is not anecdotal.
In my opinion, it is essential for any AI user to know how to use it correctly in all circumstances. There are training courses available to make everyone aware of how to use AI correctly. It is also possible, and surely the best solution, to install internal AIs. This allows the system to be used to its full potential without fear of data leakage. This is something we at 5 Degrés can help you with .
The future of Developers in the face of AI: a changing profession
I'm going to come back to the main question of my article: what is the future of Developers with the rise of AI? It's a good tool, very useful on a daily basis. Our job is often split into two parts: the development of new applications and project management. Generally speaking, our customers need support on development issues rather than on the code itself.
Conclusion
The feedback I can give is that, at present, AI is no substitute for hard-won experience. It will indeed be able to list the most widely used technologies on the market, but it's only when you're actually in the field that you'll realize the relevance of using this or that software for your needs.
Perhaps one day, we'll stop coding and only produce specs that can be understood by AI. But after all, with our computer languages, isn't that what we're already doing to a lesser extent?
So my opinion of AI has softened, and I see all the potential that this extraordinary tool will soon become - and already is.
I could have written my article with an AI in five minutes. It would indeed have saved me time. But I remain convinced that the difference in terms of quality is still there, and that work done with passion is appreciated.
Manon Le Guet
Consultant Python Developer
