When Governments Can't Keep Up With AI
Kyriakos Mitsotakis just said out loud what leaders privately know: no government is actually ready for what AI will do next. The real gap isn't between regulation and innovation—it's between deployment speed and institutional capacity.
A Leader Admits What Others Won't Say
<cite index="3-1">Greek Prime Minister Kyriakos Mitsotakis admitted in San Francisco this week that no government is ready for what AI is about to do.</cite> This wasn't a scripted policy speech. <cite index="3-6">In front of roughly 250 founders, investors and operators, he said openly that he doesn't have answers to many of the AI questions leaders around the world are more privately debating.</cite>
What makes this moment significant is the candor. Most heads of state on trade missions stick to pitching their countries as investment destinations. Mitsotakis went further—he essentially declared that governments are outpaced, and the public needs to understand that reality.
The Evidence Is Overwhelming
This isn't just one leader's opinion. The data backs him up. <cite index="28-2">AI capability is still accelerating, adoption is spreading at historic speed, and the systems, institutions, and policy meant to manage the fallout are lagging badly.</cite> The structural problem is clear: <cite index="24-4,24-5">governance lags partly because AI is easy to try—you do not need a full project plan to paste text into a model and see what happens, or a six-month roadmap to wire an API into a small internal tool.</cite>
Across OECD governments, the disconnect is measurable. <cite index="14-1,14-2">While 32 of 36 countries (89%) report training programmes to support AI in government, fewer offer training on using AI in public services or policymaking (13 of 36 countries, or 36%, each).</cite> Translation: governments have broad AI awareness but lack the specific skills to deploy it responsibly.
Where The Execution Gap Hurts Most
The real damage happens when governments move faster than their infrastructure allows. <cite index="21-6,21-7">Government organizations demonstrate surprisingly strong overall AI maturity, yet public sector investments in trustworthy AI technology and governance often lag behind, suggesting agencies may be deploying advanced AI built upon a shaky data foundation, increasing the risk of biased outcomes, security breaches and costly operational failures.</cite>
Consider the practical bottleneck: <cite index="14-11">only 21 of 36 OECD countries (58%) provide central support for procuring AI goods and services.</cite> When procurement readiness is absent, agencies either buy the wrong tools or skip buying entirely and build fragile workarounds.
Mitsotakis himself identified this problem earlier this year. <cite index="2-1,2-2">He noted that technology is advancing at extraordinary speeds, but too often public institutions are operating on an outdated operating system and rules, and if we want AI to serve society, governments must significantly update their own software.</cite>
Why This Matters for Developers and Teams
If your organization builds for government or works with government systems, this gap directly affects you. <cite index="27-1,27-7">The United States still has no comprehensive federal law governing artificial intelligence, and a new analysis describes a structural mismatch where frontier AI systems now capable of autonomous, offensive operations against real-world targets are advancing faster than the institutions meant to govern them.</cite>
That means compliance requirements are shifting mid-project. Procurement processes that should take months drag on. And the rules you're supposed to follow today might change next month.
For teams building with AI tools—whether you're using LLMs from OpenAI, Claude, or open-source models—understand that regulatory clarity is coming, but it's coming late. <cite index="22-3,22-4,22-5">Innovation evolves at machine speed, while governance moves at human speed—as AI adoption grows exponentially, regulation is lagging behind, and worldwide, governments are scrambling to regulate AI with fragmented and uneven approaches abounding.</cite>
What To Do Right Now
If you're deploying AI in production, don't wait for policy to catch up. Build with governance in mind: document your data sources, log model outputs, and implement human review checkpoints for high-stakes decisions. Tools like NeonCodex AI can help you test AI workflows before shipping them to users—it lets you validate behavior across different scenarios and catch governance issues early.
For policy people and procurement teams: Mitsotakis' candor is a permission slip to stop pretending you have certainty. Instead, build incremental pilots with clear success metrics, invest in upskilling your existing teams, and treat vendor evaluation as an ongoing process, not a one-time purchase decision.
The gap between where AI is and where governance can reach won't close fast. Plan for that reality, not the fantasy.
Source: [TechCrunch](https://techcrunch.com/2026/09/22/were-already-fighting-yesterdays-battle-greeces-prime-minister-gets-candid-about-ai/)