We are currently experiencing a revolution in software development. Artificial intelligence is writing the code. Will humans soon be obsolete in software development? On the contrary, they are becoming increasingly important.
In spring 2025, the founder of Anthropic, Dario Amodei, predicted in an interview with the American Think Tank Council on Foreign Relations that 90 per cent of code will be written by AI in just three to six months. Amodei is predicting a revolution in software development. What will change as a result? The work of software developers and everyone who works with software code. They will do less programming themselves. They will check, test and adapt more. In future, AI tools will write the book. The software developers will do the editing. Humans will remain extremely relevant in this changed distribution of roles. We humans must keep track of things when AI is generating code at a rapid pace. We have to be alert. Perhaps more than ever. We need to be able to understand what the individual code elements mean and how they interact with each other. We need to stretch our own minds and critically review the code that AI generates. AI generates the code. We still bear the responsibility.
We humans must keep track of things when AI is generating code at a rapid pace. We have to be alert. Perhaps more than ever.
Will the job profile of the software developer be a relic of the past? Not at all. First, you have to understand that software development is more than just coding. Coding is a crucial step on the entire journey in the software development process – no question about it. But coding is only the craft. Software development encompasses both technical and engineering aspects. In my experience, human developers need countless hours of practical work to master the craft of coding. The much more complex task in the software development process is to understand what the customer wants, what problem to solve with code, how to structure the architecture, what design to choose for the optimal result, how the individual code components interact with each other and with the surrounding IT system, and ultimately to manage the project and communication with all stakeholders of the project. This sometimes requires interpersonal sensitivity, a broad range of experience and contextual knowledge, creativity and foresight. Skills that are being developed more and more in AI systems, but also qualities that characterize experienced and competent software developers.
I have a vision: in the future, human software developers and AI systems will work hand in hand in the software development process. They complement each other, give each other impulses, scrutinize each other and develop the best possible software for the customer in this fruitful and effective duality.
For some months now, T-Systems has been using a solution that supports software developers in coding: the AI Engineer. The AI Engineer automatically generates, documents and tests code. The solution is fluent in current programming languages such as Python or Typescript. The AI Engineer can also handle older code in Cobol or C++ and helps to translate it into a desired modern programming language. Applications that used to take months to develop can now be programmed by the AI Engineer within minutes. For example, an app that issues temporary virtual car keys to friends and displays all of the car’s features and fill levels on a dashboard. Or a digital citizens’ office where you can apply for ID cards and notify a change of residence. A simple prompt that paraphrases the desired specifications is enough. The AI Engineer does the rest. It generates and translates code, documents, structures and tests it so that it can be reproduced in the subsequent software development process and adapted in modules. The developers manage the AI Engineer, carry out quality control and check how the code can be integrated into the customer’s existing IT infrastructure.
The AI Engineer reduces the development process from six months to just a few minutes – including testing and deployment. Simply with a short natural language prompt. As part of our AI Foundation Services, the AI Engineer exemplifies the technological leap we’re witnessing in AI, cutting costs and helping our clients unlock their full business potential.
By using the AI Engineer, the approximately 1,100 developers at T-Systems will have more time for more demanding tasks that require more creativity or contextual knowledge. The AI Engineer will enable a significant increase in productivity, high privacy and security standards and reduce software development costs at the same time.
When using the AI Engineer, the responsibility remains with the software developer. They have the tool in their hands and control how and for what purpose the code is generated. The AI Engineer was developed in accordance with the highest security and privacy standards and has undergone an extensive privacy and security assessment. It also fulfils the high requirements that Deutsche Telekom places on its AI solutions and is compliant with the EU AI Act.
The AI Engineer is part of T-Systems’ AI Foundation Services, which provide a secure environment for the development, testing and production of AI models and help companies to scale AI solutions efficiently and securely.
