Are we all overestimating AI?

Monday, July 1, 2024 Posted by bloggerdaddy

Monday, July 1, 2024

Welcome to TechCrunch AM! It's been a busy weekend in Tech Land, where business and politics intersect. In the U.S., that looks like a Supreme Court ruling with ripple effects on how industry is regulated, and in Europe, it looks like Meta potentially paying steep fines. We've also got notes on what genAI can't do, how to counter AI scams, more money for HR tech, and much more. To work! — Rebecca

TechCrunch Top 3

Image Credits: Rudy Sulgan / Getty Images

1. One more decision to skew the balance of power: By overturning Chevron v. Natural Resources Defense Council, the U.S. Supreme Court has granted itself exclusive power currently exercised by regulatory agencies. How does this relate to tech? The 1984 Chevron decision established that courts must defer to independent agencies as subject matter experts where the law is ambiguous. This ruling may, in effect, give a lot of power to industry lobbying groups. Read More

2. Generative AI, shmenerative AI: Rodney Brooks, an MIT robotics pioneer, thinks people are vastly overestimating the promise of generative AI. Brooks says when a human sees an AI system perform a task, they tend to think it can do way more, and thus tend to be overly optimistic. Read More

3. EU says Meta's "pay or consent" model doesn't fly: An investigation by the European Commission has found that Meta's binary offer to Instagram and Facebook users — either pay to avoid ads, or consent to tracking and see targeted ads — doesn't comply with the Digital Markets Act. Meta, of course, will fight that tooth and nail, but this is still good news for privacy advocates. Read More

SPONSORED BY BANK OF AMERICA

Advice from top tech investors

Read their insights for tech company growth from Bank of America's Triangle Innovation Summit.

Get Insights

Morning Must Reads

Image Credits: tadamichi / Getty Images

HR tech is low-key raking in cash: While we all get distracted by AI, HR tech startups are quietly making and raising a ton of money: Japan's SmartHR just raised a $140 million Series E to continue building its product, which it says compares to Rippling's own suite. SmartHR's got the receipts to prove it too: Its ARR touched $100 million as of February. Read More

Don't get scammed, yo: The advent of generative AI has led to a proliferation of scams. Things like cloning your friends or family members' voice to get you to send money, personalized email phishing, or identity fraud. Want to educate yourself on the risks and fight back? Here's a handy guide: Read More

Stranded? Who? Me?! No Way! The two astronauts who were brought to the ISS on board the Startliner are going to stay in space for a few more weeks, but NASA and Boeing officials say they're definitely NOT stranded. They're just chillin' out, getting up close and personal with the capsule before it returns to Earth, for…research purposes… After last week's debate, can you blame them? Read More

Detroit Police has limited its use of facial recognition technology as part of a legal settlement, and they can't arrest people based solely on the results of a facial recognition search. There are other details to the new policies, but the ACLU is calling it a win, saying other police departments should use facial recognition technology similarly, especially since it's been known to misidentify women and people of color at higher rates. Read More

Decentralized finance meets centralized tax reporting: Crypto platforms will now need to report transactions to the IRS starting in 2026, per new regulations. Decentralized platforms that don't hold assets themselves will be exempt. "We need to make sure digital assets are not used to hide taxable income," the IRS said. Read More

Helping companies use AI responsibly: In the latest issue of TechCrunch's series highlighting women doing amazing things in AI, Dominic-Madori Davis interviewed Sarah Bitamazire, chief policy officer at boutique advisory firm Lumiera, where she works with high-level decision-makers to help them implement AI properly and responsibly. Read More

1 billion stolen records and rising: We're only halfway through 2024 and we've already seen some of the worst, most damaging data breaches in recent history. Zack Whittaker has a roundup of the biggest cybersecurity incidents of 2024, how they happened, their impact, and what could have been done to prevent them. Read More

Around the Web

Want to restore a forest? Just ask the birds: Researchers from Canberra have apparently used AI to figure out what kinds of trees they should work on restoring, and they went about it by studying those who live in trees the most: birds. Using AI, the researchers studied trees and figured out how to build structures that the birds prefer, not installments that look like trees to our human eyes. The full study's on The Conversation, and you should definitely read it. Read More

You might have to eat moss on Mars: Forget potatoes, researchers say they've found a desert moss that can survive on Mars that could be pivotal to establishing life on the planet, The Guardian reports. The moss is found in regions like Antarctica and the Mojave desert and can handle drought, high levels of radiation and extreme cold. Read More

The era of AI-washing: We've already established that companies are probably making over-inflated claims about their use of AI, but now we have a fun new term for it, per the BBC: AI-washing. This is becoming more salient in the era of genAI as companies claim to use AI when they're actually using much less sophisticated tech. Read More

Before You Go

Image Credits: Ageless Innovations/Joy for All

Everyone deserves a furry pal: Furry robotic pets have seen a lot of traction in countries with aging populations like Japan. Now, New York State is working to embrace robotic critters. The state has sent out thousands of robot cats, dogs and birds produced by Ageless Innovations to older adults to keep them engaged and help fight the "epidemic of loneliness." Read More