You can skip the line for a sold-out gym class with an AI agent today. What else can you skip?

Today, AI seems the most lucrative (yet to be seen profitable) industry, where the majority of the world's economies are shaping their future strategic economic, financial and structural builds. In fact if you have the tools and a great pitch deck, you can receive a fairly chunky backing from an investor or an EU institution within Europe. With such crushing demand for innovation, the EU Commission itself also just opened a call to unlock more than €30 billion in AI infrastructure investment (European Commission, AI Gigafactories call, 30 July 2026).

Yet every (profitable) industry we trust with something important is regulated. From food to Airlines to Public transport, Energy, Financial services, Banking, Law Enforcement, the press, Healthcare to Fitness supplements or services and Data compliance... to name a few. And yet the one technology we are currently investing our pensions in, building our livelihoods around and handing our collective intellect over to, has only just gained anything resembling a real regulatory framework. Yet, it's a framework still being watered down as it arrives...

The EU AI Act is the world's first comprehensive AI law (artificialintelligenceact.eu) and its high-risk obligations only became enforceable on 2 August 2026 (European Commission, press release, 31 July 2026) just weeks before piece was pulled together. Outside the EU, the picture is thinner still: the US has favoured a far lighter-touch, with a patchwork of state laws filling the gaps (Collibra, AI regulatory compliance in 2026). Even within the EU, the rules are being softened almost as fast as they're being written: carve-outs for industrial AI, extended simplifications for mid-sized companies and a delayed rollout of regulatory sandboxes were all agreed in the months before enforcement even began (Latham & Watkins, May 2026; Council of the EU, 7 May 2026).

Which means, until only very recently, anyone that labelled their tech as 'innovative AI' could get both backing and the room to build something (forget all the other industries driving our economies and in much need of investment - not just for jobs yet for robust continued and structural functionality). This makes it far easier for those not thinking of the long-term impact yet short term profitable gains to, for example, design a pathogen, a piece of malicious code or a tool built to deceive at scale. And it's so accessible now, you can even skip the line now to grab your spot at a sold-out gym class simply by training your AI agent to do so. This means all societal rules are up for grabs to be re-written and re-patched - in a continuous, disruptive loop.

If the new rules arrive fast and firm enough to actually change that in practice is still an open question. That's why we're launching the AI Ethics Table to help all founders working in AI avoid costly mistakes that will only show up years down the line and if you think 'nay, why bother, I'll be retired by then', just take a backwards glance at the impact of the fossil fuel, fast-fashion or the plastics industries or the aftermath of the European summer fires and you'll want a seat at that Table early on.