☕ TL;DR (because you’ve got 30 seconds and a coffee going cold)

• The EU just opened an antitrust probe into Google for using publisher and YouTuber content to train Gemini, without paying or asking. This could rewrite the rules for everyone in AI.

• Elsewhere, 25% of teens are calling chatbots their “friends” and leaning on them for mental health help (nope, not kidding).

• The UK’s warning that foreign powers are deepfaking their way into undermining support for Ukraine, and over 200 UK politicians want to regulate advanced AI systems now.

• Finally, environmental groups are pushing for a moratorium on U.S. data centers because the AI industry is turning into a power-hungry monster.

📌 Reality is finally catching up with the AI hype. Regulators are on the move, safety cracks are wide open, and the infrastructure bill is here—and it ain’t cheap.

Hey founder, 💁🏻‍♀️

Is your business model actually sustainable?

Find out for free with the agent I built:

👉 Learn how to earn $10 in Impressive credits

Want to hang out with other Impressive builders and founders?
👉 Hop into our Discord

1️⃣ EU opens fire on Google’s training tactics 🇪🇺

The European Commission has officially launched a competition investigation into how Google is using online content from publishers and YouTube creators to train Gemini—without proper compensation or a “would you mind if...?” 😂

What’s at stake?

Potentially the entire data backbone of generative AI. If this ends with a “you can’t do that”, every AI company that built on scraped content (yes, all of them) is now exposed.

This could flip the entire AI business model on its head.

Suddenly, everyone needs licensing deals.

Training costs could go from “just scrape it” to “millions in IP contracts”.

That’s a massive shift in growth economics and a deathblow to early-stage companies who can’t afford to play in that league.

OpenAI, Anthropic, and the rest?

They’re watching with popcorn... and heartburn. 🍿👀

2️⃣ 1 in 4 teens are now using AI chatbots for therapy. They call them “friends” 🧠

A new study shows that roughly 25% of teenagers are using AI chatbots for mental health support. Some even describe the bots as “friends”.

Yeah. That’s where we are now.

Let’s be real: teens turning to AI exposes that human support systems are breaking down, and while it might sound sweet, it’s a massive red flag for safety.

These chatbots aren’t trained therapists.

They don’t detect crisis signals.

They can’t escalate when things get dangerous.

And they’re running without a shred of regulation.

💡 Growth without guardrails doesn’t scale, and if you’re building in this space prepare for lawmakers, regulators, and parents coming at you from all sides.

3️⃣ Deepfakes go geopolitical; support for Ukraine in the crosshairs 🎥

The UK’s Home Secretary, Yvette Cooper, just confirmed what many suspected: foreign governments are using AI-generated videos to spread disinformation and erode public support for Ukraine.

This isn’t experimental; it’s operational and… terrifying.

Deepfake propaganda is now standard statecraft. The days of wondering “what’s real?” are over. Welcome to the era of total narrative manipulation at scale.

For AI founders and builders? This isn’t just a moral issue. It’s a growth inhibitor. Expect calls for watermarking, detection layers, content filters, and serious compliance burdens baked into whatever you’re building.

4️⃣ UK politicians join forces: “Regulate this AI thing. Like, now.” ⚖️

More than 200 UK parliamentarians, across party lines, are demanding regulation for powerful AI systems. Cross-party alignment like this is rare. When it happens, it usually means legislation isn’t far behind.

This could mark the UK’s pivot toward an AI Act-style framework, similar to the EU’s. If so, companies operating there (or expanding into Europe) now face a new regulatory landscape.

→ One more layer of complexity.

→ One more set of hoops to jump through.

From a growth strategy perspective, the fragmentation of AI laws (US states, EU, UK) means your compliance stack better scale as fast as your product does. 🫠

5️⃣ AI’s climate bill just arrived: 200+ groups demand data center freeze 🌍

Over 200 environmental organizations have called for a moratorium on new U.S. data center construction. Their reason? AI is gobbling up energy like it’s going out of style.

This is the first real environmental pushback against AI infrastructure, and it’s gaining steam.

If regulators side with the climate camp, new data centers will slow down, and if that happens, so does AI innovation.

You can’t train frontier models without compute, and you can’t scale compute without hardware. Say hello to the physical limits of exponential growth.

If you’re building AI products, consider this your early warning: clean growth may become the only acceptable path forward.

Enjoying What Just Happened in AI? Buy me a coffee :)

Buy Laura an extra boost

💡 Smart Takeaway

Today’s stories all point to one thing: the AI party is hitting the real-world hangover.

• The EU is investigating training data (read: expensive legal exposure).

• Teens are using AI for therapy (and we’re not ready for the liability).

• Foreign states are weaponizing deepfakes (truth is losing).

• The UK wants to regulate AI (and they’re serious this time).

• Environmentalists are drawing red lines (and they’ve got momentum).

Growth, from now on, will be linked to your ability to navigate legal landmines, build trust with users, and prove your product doesn’t wreck the planet.

Founders and leaders who still think this is a Wild Westscape? 🤠

They’re about to get trampled. The new game is regulation-ready, safety-verified, and infrastructure-conscious.

Build accordingly.

🔍 Bonus Insight

If you only remember one thing from today, make it this: the EU’s investigation into Google could rewrite the economics of generative AI.

If regulators decide that scraping content without paying is anti-competitive, or worse, illegal, then retroactive licensing becomes a very real (and very expensive) thing. And the companies that built billion-dollar valuations on the idea that “training data is free” will have to face the music.

→ We’re talking billions in unaccounted costs.

→ Profitability models stretched thin.

→ IPO timelines pushed back.

📧 Have thoughts? Rants? Hot takes? Hit reply. I read everything.

Leave a comment