Europe's AI Regulation Strategy: Pioneering or Pigeonholing Innovation?

This is The Startup Breakdown, the newsletter where we learn, laugh, and love startups. By joining this growing community of hundreds of future startup aficionados (think i spelled that right?), you're getting a beachside view of the ocean that is the startup and VC scene. This ain’t your grandpa’s newsletter, so prepare yourself for an inbox full of 4/20 jokes and Succession references.

Happy Wednesday, folks.

Full transparency, growth has plateaued a bit. Part of that is the self-imposed purge automation which gets rid of subscribers who don’t engage (open & click) on newsletters, but I also find that I have strayed a bit from listening to your feedback.

You’ve probably heard of Product-Market Fit, so here I am seeking to find Content-Audience Fit.

So here’s my ask: what do you want this newsletter to be?

It has come a long way since being me just ranting about the latest news in tech. Huge shoutout to those who were here since the Pigskin Economics days…

I started out writing for myself. Now, I have an actual audience, and I am writing for you.

So even if we have never actually interacted, I am asking you to shoot me an email (can just reply to this newsletter) telling me exactly what I can do to make this experience better for you.

I can’t tell you how much I appreciate it.

[RESPONSE REQUESTED]

What is in the EU’s AI Act?

The European Parliament just gave a Masterclass in how to remain a second-tier tech region, and they didn’t even charge for it!

The new AI Act takes a “risk”-adjusted approach to regulation, categorizing AI by its perceived level of danger, from simple content recommendation algorithms which are labeled “no risk” to the more extreme use cases such as social credit scoring and predictive policing which are each considered “unacceptable.”

This much is reasonable, as it’s clear that not every AI product presents the same level of risk and shouldn’t be blanket-regulated as such.

More risky (their words, not mine) technologies can see regulation scaling up to an outright ban, while those on the lower end of the spectrum are given a voluntary set of guidelines that may choose to follow.

Many of these rules are intended to be applied at the ground level when a product is released, but the bloc is coming after established players, too, many of whom aren’t even “AI” companies.

While you were pounding back Guiness for some guy named Pat, first-years at Instagram, Snapchat, and Twitter (prolly Grok actually) were 4 coffees deep pumping out some BS statement on the comprehensive policies that their employers are taking to curb AI-generated content.

Other restrictions are even more cumbersome for AI companies looking to operate in the European market, namely the requirement to comply with EU Copyright Laws (looking at you Sam…) and disclose exactly where that data came from (looking at you Mira…)

Europe is a leader in climate-friendly policies (though this is easy when your economy is far smaller and less reliant on traditional fuel sources), and aligning with that prerogative, the new AI Act will also require companies to disclose their climate impacts. Spoiler alert, they won’t be good.

The EU has long been 10 steps behind in the race for tech supremacy. There’s a reason one street in SF has more unicorns than the entire Union combined.

One reason for their obsession with being the first to craft policy on anything and everything. You could stub your toe and they’d try to ban tables with legs.

In many of these cases, their hope is that they’re setting a precedent that others follow, and this time is no different. Imagine the smug look on those Belgian faces when Biden turns his own AI thoughts into bills.

Last week also saw the final report on an American government-sponsored analysis of the threats of AI, and it was arguably even more restrictive on companies than the EU’s, recommending many similar policies as well as a government-imposed cap on the power used to train models…

Notably, two European countries were against the final version of the law, namely France and Germany. They can also be referred to as the only real national economies in Europe.

They feared the dampening of dynamism that the law could, and likely will, impose.

The law will not prevent AI technologies from developing. Instead, the chains of bureaucratic compliance will merely slow the already lagging pace of development, preventing the continent from taking advantage of a generational technology that, if cultivated correctly, could close any gap between it and the more dynamic economies of the world.

Nothing more than a whole lotta headaches and paper cuts.

Such restrictive measures will not happen in the US. For better or worse, we are governed by big tech, and they have too much at stake to let something like these administrative burdens weigh them down.

As for what it means for startups, if you’re trying to convince yourself that there’s a better place to build than the US, just stop. Get your ass to the States and build the future.

Startups may also have to consider just how important the European market is as a consumer base. Will your growth truly be elevated enough in Italy or Greece to warrant hiring an entire legal department to handle compliance?

For most, the answer is probably no.

The Atlantic Ocean might as well have doubled in size given the separation between the US and Europe. However, the Pacific might be just a creek given Asia’s pace of development and the very eager government encouragement they’re receiving.

Europe hamstrung itself. Now, the race for AI gold is neck and neck between Altman and Xi.

The European Parliament's recent AI Act introduces a risk-based regulatory approach for AI technologies, sparking debate over its potential to stifle innovation in the tech sector while aiming to set global regulatory standards.

Have other feedback? Reply directly to this email and let me know!

Cheers to another day,

Trey