Ethical AI Is an Engineering Problem

Interview:

What is your session about, and why is it important for senior software developers?

This session explores why ethical AI is not only a policy or governance issue, but also a core engineering challenge. Many of the risks associated with AI systems—bias, lack of transparency, security vulnerabilities, or excessive computational cost—often emerge from technical design choices made during development.

For senior software developers and architects, this means that ethical outcomes are not abstract ideals but properties of the systems they design. The talk examines real-world failures in AI systems and shows how architectural decisions, data pipelines, model evaluation, and system observability directly influence fairness, transparency, security, and sustainability.

Technology often evolves much faster than regulation, meaning engineers frequently operate in areas where clear rules or standards do not yet exist and new risks are still emerging. In this context, ethical principles such as fairness, transparency, security, and sustainability can serve as practical guides for designing and implementing new technological products responsibly.

The goal is to help engineers understand how ethical principles can be translated into concrete engineering practices across the AI lifecycle.

Why is it critical for software leaders to focus on this topic right now, as we head into 2026?

AI systems are rapidly moving from experimental tools into components of critical infrastructure embedded in products, platforms, and decision-making processes. As adoption accelerates, the consequences of poorly designed AI systems are becoming more visible—from discriminatory outcomes to security vulnerabilities and unsustainable computational costs.

We are at a stage similar to other major technological transitions in history, where new engineering practices must emerge to make systems safe and reliable at scale. Software leaders play a key role in shaping those practices. As we head into 2026, organizations that treat ethical AI as an engineering discipline—rather than an afterthought—will be better positioned to build trustworthy, resilient, and at the end of the day better AI systems.

What are the common challenges developers and architects face in this area?

One of the biggest challenges is that many ethical risks in AI systems are difficult to detect using traditional software engineering approaches. Bias may originate in training data, explainability can be limited by model architecture, and security vulnerabilities can arise from new attack vectors such as prompt injection or model extraction.

Another challenge is translating high-level principles—such as fairness or transparency—into concrete engineering practices. Teams often lack clear architectural patterns, evaluation metrics, or operational tools to implement these principles throughout the AI lifecycle. As a result, ethical considerations are frequently addressed too late, when systems are already in production.

What's one thing you hope attendees will implement immediately after your talk?

I hope attendees start treating ethical properties of AI systems the same way we treat reliability, performance, or security—as engineering requirements that must be designed, measured, and continuously monitored.

In practice, this means incorporating fairness evaluation, explainability checks, security testing, and resource efficiency considerations directly into the development lifecycle. By embedding these practices early in system design and architecture, teams can build AI systems that are not only powerful, but also trustworthy and responsible.

I also hope people leave with a stronger curiosity about the risks and limitations of AI: following the growing body of research in the field, questioning assumptions, and becoming more aware of the societal and technical implications of the systems they build.

What makes QCon stand out as a conference for senior software professionals?

QCon stands out because it focuses on real-world engineering challenges and lessons learned from building systems at scale. The conference brings together experienced practitioners who share practical insights rather than theoretical ideas, creating an environment where senior engineers and architects can learn directly from peers facing similar problems.

This emphasis on practitioner-driven knowledge makes QCon particularly valuable for professionals responsible for designing and operating complex systems, especially as new technologies like AI reshape the software landscape.