IndustryWorking in IT and reading the news can be rather scary these days. Many experts fear significant job cuts in a sector that felt so safe for several decades, with growing income and job security.
Gen AI, agents, and bots seem to overwhelm humans, leaving us with pictures in our minds of dystopian sci-fi movies like Terminator, at least for the older generation. IT was seen as the holy grail, absorbing labour and permanently reporting a lack of qualified employees all around the world.
Today, we see hourly rates for craftsmen like carpenters or mechanics that are way higher than those of experts in the IT sector. Doomsday ahead!
STOP!
Before we all turn depressive and ask for psychiatric medication, let’s do a reality check. AI is here to stay. But it is still young and learning how to walk. Some of its capabilities are impressive, but far away from being perfect. For the foreseeable future, Gen AI still requires checks of its results and control over what it does.
So, are developers doomed to extinction? The answer is no, for two reasons. First, as AI is still far from real imagination and creativity, human developers bring the right grain of salt. Second, we still need experienced engineers to establish the guardrails, differentiating poor code from good code. We learn from our experience and the mistakes we have made in the past.
It might be seductive for enterprises to cut staff and save costs. But there will be organizations achieving more with the same team size, gaining a competitive advantage and putting pressure on competitors through a higher rate of innovation.
What is good for the economy is unfortunately available to criminal minds, too. The ongoing digitalization of our lives opens new points for attacks. As Gen AI lacks ethical preferences (so far), it can also be used to accelerate the creation of malware, vulnerable components, and the exploitation of known weaknesses in deployed software.
This is a race that humans cannot win. The examination and qualification of code source or executable in a traditional, manual way will not allow us to keep up with the momentum of new code being created every minute.
Every coin has two sides; we have and need AI to support our well-qualified experts in assessing risk and deciding on actions in real time. Point solutions that address some of the challenges are a nice step forward, but the upcoming avalanche of unqualified code requires end-to-end coverage.
It starts with an ongoing evaluation of source code on the developer machine and within the IDE to detect questionable patterns, providing immediate feedback. Centralized SAST solutions keep control over committed items and are ready to be injected into the CI/CD pipeline, including SCA and compliance observation. Binary management helps us block malicious components, manage SCA and SBOM, and trace exactly what went into production. It also enables fast traceback in case flaky software is detected in production.
BTW, still often neglected but increasingly important is the lifecycle management of LLMs and agents, similar to the code and applications we use to create them.
A final aspect is the management of risk introduced by Gen AI using human identities while doing the work. Do you know exactly what the agents can do and what risks they contribute? You might argue that you have parts of this in place. Let’s try an experiment, randomly remove words from the previous section. Does it provide the same level of information, or is something missing?
End-to-end is not optional, it is a must-have. Leaving parts of the toolchain uncovered causes significant risk for your organization.
2026 is the time to invest in the future of your software development. We encourage you to contact us for your free risk assessment, providing clear information for investing your budget smartly and driving productivity, quality, and security.
ASERVO Software GmbH
Konrad-Zuse-Platz 8
81829 München Germany
Tel: +49 89 7167182 – 40
Fax: +49 89 7167182 – 55
E-Mail: Kontakt@aservo.com
Copyright © 2023. ASERVO SOFTWARE GMBH