We spend hours finding elegant solutions to nicely stated problems. In competitive programming, problems come tied with restrictions, inputs, and outputs. But when you're dealing with artificial intelligence, the “problem statement” isn’t even written yet — and that is riskier than any unsolved Div. 1 problem on Codeforces.

The Unspoken Aspect of Progress in AI

The majority of the talk depicts AI as absolute progress: faster, better, smarter. But the real world is messier. Large models require massive power, sometimes equivalent to the annual electricity of countries. Around the same time, the same models are sold as climate optimizers and solutions. This “AI paradox” — technology both problem and solution — forces us to ask the question: do we really build responsibly?

Bias is another iceberg beneath the surface. If training data reflects social inequalities, models will replicate them. Generative systems already spread misinformation at scale, threatening trust in journalism, politics, and education. Without transparency, AI becomes a black box that even its developers can’t fully explain. In a competitive programming analogy, it’s like submitting code that passes tests but whose logic you don’t understand — except the stakes here are real lives and institutions.

Why Governance Matters

I initially considered the regulation of AI to be a straightforward problem: the government just needs to enact new legislation. However, listening to the views of activists, lawyers, and industry players led me to the conclusion that the regulation is not a one-off solution. Businesses move forward into the gray areas where there is little regulation, and society resists. Efficient governance entails dynamic, open inter-sector collaboration. It is relevant to us as future computer programmers because we will be the coders of the “invisible routines” beneath these systems. If we do not take transparency, use of energy, and fairness into account when we design the systems, nothing done retrospectively will quite fix the damage.

Why Programmers Should Care

Construction, healthcare, law, education — every sector is adopting AI. But adoption isn’t automatic. Engineers want usability, regulators want compliance, and business leaders want ROI. This means technical brilliance isn’t enough. If we ignore the human side of AI systems, we risk building tools that are either mistrusted, misused, or outright rejected.

As programmers, we are used to debugging performance bottlenecks. AI adoption has its own “bottlenecks” — trust, regulation, and sustainability. Understanding these is just as important as optimizing runtime.

Call to Action

The competitive programming mindset — clarity, precision, creativity under constraints — is exactly what we need for AI governance. We need to treat transparency as a required constraint, fairness as part of the problem statement, and sustainability as a non-negotiable limit on complexity.

AI will not replace human problem-solving. But if built responsibly, it can become a partner that empowers rather than undermines us. As programmers, we should start asking not just: “Can it be built?” but also: “Should it be built this way?” That mindset shift may be the most important algorithm of all.