AI regulation FINALLY here

There's a number called "Graham's number" that's so mind-blowingly huge, that if you even tried to cram it into your brain, your head would collapse into a black hole due to the information's mass. Seriously.

It's involved in a problem in Ramsey theory, and mathematicians have worked out that the answer is between 13 and Graham’s number. Talk about being vague…

TODAY:

The EU is laying down the law on AI.

AND AI and misinformation, a neverending story.

And of course, a BIG SHOUTOUT to our partner, Shortform. Their new AI-powered Chrome Extension is quickly becoming our most useful.

AI REGULATIONS

EU FINALLY passes the world’s first AI regulation

The European Parliament has finally green-lit the world’s first extensive AI law, receiving a vote of 523 in favour, with some debate on its economic impact.

Set to kick off this May, this AI Act brings a risk-based method to the table, classifying AI by potential hazards and setting rules accordingly.

Here’s what the key highlights include:

• Bans: No-go for AI that tweaks behaviour, does social scoring, or uses biometric data without permission.

• High-risk AI: Think AI in vital areas like health, education and law enforcement, needing tight checks before and after hitting the market.

• Generative AI: Tools like ChatGPT must be open about how they work and the data they learn from, respecting EU copyright systems.

The act is detailed, covering from spam filters to complex systems, ensuring transparency and public oversight.

Coming Soon!

It’s set to fully rollout by mid-2024, with banned practices ending within six months and a complete enactment within two years.

Breaking these rules could lead to fines of up to 35 Million Euros or 7% of a global yearly turnover.

Moreover, the Act’s provisions mean AI models must detail their training data, a step towards respecting content creator’s rights.

While small businesses get some leniency, the core message is clear: AI development must not infringe on individual rights or freedoms.

This legislation not only sets a precedent in AI regulation but also poses a challenge to global tech governance, making a pivotal moment for the tech industry’s future.

Our View:

Sounds like a LOT of paperwork…

Vote for live results and see results + opinions from yesterday at the bottom of the email.

PRESENTED BY SHORTFORM

Summarise EVERYTHING with Shortform

Shortform is so much more than just book summaries. Their brand new AI-powered extension lets you apply the power of Shortform’s summarisation to any webpage.

With it, I’m able to take in so much more from the web, without wasting any time - my most valuable resource!

Summarise at will!

Whether you’re checking the latest news, or browsing an epic LinkedIn post - Shortform can pull all the key insights into a handy summary inside your web browser.

Unlock:

Instant Summaries: One click, and boom—you get the core ideas from any content online, fast.

Universal Support: It works on anything—YouTube videos, news articles, and long emails from your boss. If it's got text, they've got you covered.

No extra costs: Costs nothing on top of your Shortform membership!

For most of us, time is our greatest asset and should not be wasted. With Shortform, you can dive into the core of any online content without drowning in it. Gain insights, save time, and stay informed, effortlessly!

That’s the power of Shortform - not only can you read nonfiction bestsellers in 90% less time, but now you also get the power to summarise anything!

Want to give the Shortform AI extension a whirl?

You can grab a Shortform free trial - but Mindstream readers also get 20% off all annual plans!

Try Shortform’s AI today! →

AI Image or Real

Every Saturday we get you to vote on an image and guess whether it’s a real photo or an AI-generated image.

The hallway of Trinity College’s library. But we don’t know if it’s AI or real!

Can you tell? Vote below!

The answer and results are revealed in tomorrow’s Mindstream!

AI MISINFORMATION

Is AI really that bad?

AI is changing the game in many fields, sparking debates on its pros and cons.

The buzz around ChatGPT’s Sora has many people questioning whether AI is good or bad for society.

One major worry is about AI helping spread false information through fake images, texts, and videos, making it easier to deceive people.

But, there's a bigger issue at hand. It's not just about the AI; it's about the system it operates in.

Our current setup already allows misinformation to spread easily, with or without AI's help.

For example, a fake story about an explosion near the Pentagon, pushed by an AI-generated image, really shows the cracks in how we check and understand news, rather than pointing the finger at AI alone.

Not all misinformation needs high-tech AI

Old-school tricks using simple photo editing tools are still around and effective.

This hints at a deeper problem: our digital environment has weak spots that can be misused, and it's not always about AI being super advanced.

Focusing too much on AI and misinformation might also take the heat off big tech companies that play a big role in spreading fake news.

The issue is more about how misinformation is spread through these platforms because of policy loopholes, rather than the AI-generated content itself.

As big elections come up, how big tech firms handle fake news and political ads is under the spotlight.

The real challenge is not just about stopping fake news but about dealing with the narratives it creates and how people react to it.

In short, the main problem isn't AI itself but how we handle the misinformation it can help create.

Our View:

We all know our uncles’ Facebook posts are about as reliable as a used car salesman with a toupee, but AI isn’t helping either.

Mindstream Picks

The upcoming Galaxy Z Flip 6, rumoured to offer potential improvements, could be a worthy upgrade for those who skipped the Galaxy Z Flip 5 later this year.

SpaceX's Starship prototype successfully reached orbital velocity after completing a hot-stage separation during its third flight test, according to Elon Musk.

UK government urges telecoms to share existing infrastructure to ease broadband deployment concerns.

Bitcoin dipped to a one-week low amid profit-taking and concerns over U.S. inflation, dropping over 5% to $66,629.96 before recovering slightly.

CCP Games, creator of Eve Online, has made its Carbon Development Platform open source, allowing players and developers free access to the engine.

AI Art

Our Image of the Day

Daily Image Prompt

Get art inspiration by trying this prompt in your image generator of choice.

Happy with the result, or have some other artwork you’re proud of?
You can send in your artwork and we will feature it, with credit to you.

Submit your artwork to Mindstream →

We get a lot of submissions, but we do look at every single one! So please don’t hesitate to send us your art.

Yesterday’s Poll

“What would you want GPT 4.5 Turbo to have most?”

MORE SPEED! 8%

More logical thinking 51% ✅

Less hallucinations 41%

Reader’s Opinions

“GPT 4.0 is soo slow compared to the competition.”-notbot

“The ability to webscrape for information without the need to rely on expensive research services to provide this, would be ideal in an upcoming release.”- d.j.seeley

“We need more accurate information regarding specific technology markets and, so far, ChatGPT does not meet the expectations” - jcperez

Submit your opinions in our daily poll for the chance to be featured!