Artificial intelligence is accelerating software development tasks from code generation to testing. Its greater impact, however, lies in building applications that are more secure, scalable, and compliant - a shift that companies like Telefónica are already embedding into their workflows.
Reshaping the developer's day-to-day work
Within Telefónica, AI supports nearly the entire development process. It helps refine specifications, design architectures, generate code, and create tests. Developers spend less time writing code manually and more time directing, validating, and ensuring quality.
The developer's role now centers on technical responsibility: understanding requirements, translating them into well-designed solutions, and verifying integration within complex ecosystems. Testing, especially integration testing, gains importance because it confirms components behave consistently and securely.
Telefónica sees the developer's role shifting toward software architecture rather than traditional programming. AI accelerates execution, but developers supply the judgment, context, and big-picture oversight. This focus on design and validation delivers both greater speed and higher quality.
Secure development and the AI risk surface
AI only delivers real value in business applications when integrated with secure design practices throughout the software lifecycle. The need for verification doubles when AI generates or reviews code: teams must confirm not only that the solution works, but that it is secure, maintainable, and compliant.
Adopting AI also introduces new risks. OWASP's Top 10 for LLM Applications 2025 identifies specific vulnerabilities tied to AI use. This means organizations must govern and secure AI as a technology with its own attack surface.
AI acts as a capability multiplier for secure development, not a replacement for expert review. Automated scans can flag issues early, but manual review remains essential for business logic flaws and complex vulnerabilities that tools miss. Microsoft warns that AI-generated fixes must be reviewed and tested before integration.
Early vulnerability detection with AI
AI can analyze code changes, spot potential vulnerabilities, detect insecure configurations, review dependencies, and propose fixes before they reach production. Tools like GitHub combine static analysis with AI to flag pull request issues under developer supervision.
This shift toward early, automated detection helps organizations move from a reactive security posture to a proactive one. It requires clear governance frameworks and integration with existing security tools.
The skills developers now need
Fundamental technical skills - object-oriented design, APIs, databases, concurrency - remain critical. On top of that, developers must now work with intelligent assistants, assess AI-generated solutions, understand model risks, and know data protection and compliance requirements. Resources such as AI for IT & Development offer targeted training for these emerging demands.
Adaptability is key. The way software is built with AI changes constantly, so a flexible mindset and continuous learning are as important as any single tool. AI adds a layer of abstraction: developers give high-level instructions and focus on the important details, much like a workshop master.
Telefónica's research team emphasizes that one of the most valuable skills is knowing how to supervise, direct, and refine AI output, not just how to program. Developers must judge when an AI solution accelerates work and when it introduces technical debt or security weaknesses.
Telefónica's governance-first approach
Telefónica insists that AI innovation must run alongside strong security, compliance, and accountability. The company works with secure, isolated environments so data stays within corporate control and meets protection and confidentiality standards.
The approach combines AI with mature engineering practices, clear controls, and human oversight in critical processes. That combination determines whether AI adoption truly adds value.
Future trends: agents, specialized models, and on-premises AI
Over the next few years, AI development and cybersecurity will converge further. Lighter, more efficient models will embed into more workflows. On-premises models, run within corporate infrastructure, will grow in use to keep data local.
Agent-based architectures will coordinate complex tasks autonomously, and specialized models for IT security will become more common. The result won't be a replacement of developers, but a further rise in their level of abstraction, oversight, and governance responsibilities.
Why this matters for IT and development professionals
AI is handing developers and security teams the tools to work faster and more securely, but it also demands new levels of vigilance. The professional who combines technical depth with critical code review, architectural judgment, and security awareness will lead this transformation. Building those skills now - through practical experience and targeted training - will define whose career accelerates alongside AI, and who gets left behind. Courses like the AI Learning Path for Software Developers can help structure that upskilling.
Your membership also unlocks:
