Technology

➚ MOVE FAST: Getting ahead. Microsoft, BlackRock and Abu Dhabi’s MGX are launching a $30 billion fund to build AI infrastructure, The Financial Times reported and Semafor confirmed. It’s just one small piece of the massive resources needed to keep improving AI models, but it shows forethought. And it highlights the urgency that big tech companies feel on AI, as they compete with one another and with China.

➘ BREAK THINGS: Long overdue. Meta made major changes to the way it handles teen Instagram accounts, making them private by default to prevent predators from targeting young users. But for a company that coined the term “move fast and break things,” the announcement Tuesday, coming years after being heavily criticized for how it handles teen accounts, reminded people that it can move slowly when fixing things.

There are sunglasses that harness the power of AI, like the Ray-Ban Meta Smart Glasses. Now, you can buy sunglasses designed to block AI from replicating your identity or turning you into a deepfake.

Visionaries sunglasses, which I haven’t personally tested, don’t use cutting-edge electronics to throw off AI algorithms. Instead, maker NeuThroné says a simple image on the side of the glasses is enough to keep AI from using photos online to replicate your face. The catch: you have to be wearing them when the photos are taken.

The company hopes that it will find an early market with young online creators, many of whom have had their identities repurposed by AI without their consent. (Wednesday actress Jenna Ortega said she received explicit photos of herself as a child that were created with AI.)

For people like me, it’s too late. I’ve been putting photos of myself online since my 20s. These glasses highlight an interesting question for young people who haven’t plastered their social media with selfies yet or for my own kids, who I’ve kept off social media until they can decide what they want.

A radical idea to make the UAE an AI innovation hub

THE SCENE

ABU DHABI — Courtney Powell, chief operating officer of venture capital firm 500 Global, recalled an idea that a fellow investor recently floated: What if the United Arab Emirates created a “special economic zone” in which copyright doesn’t exist?

The lack of a copyright law would allow AI companies to train powerful models without worrying about lawsuits from book publishers, musicians and others who claim their work was ripped off by the technology.

Powell, who was visiting from her home in Riyadh and has helped seed a growing ecosystem of startups there, wasn’t endorsing the idea. And there’s no indication that the leaders of the UAE would entertain the concept.

The suggestion, though, highlights an increasingly common perception among some in the tech industry that the United States is coming down too hard on new developments like crypto and AI, and risks hurting innovation.

Meanwhile, Powell said the Gulf region has become an increasingly attractive place for startup founders from all over the world, from South Korea to Latin America to Russia. There were 79 venture deals in Saudi Arabia and the UAE in 2015, according to PitchBook. That number reached 402 in 2022 before dropping slightly last year to 337.

As the UAE and Saudi Arabia build out their data centers capable of training and running powerful AI models, the region could be seen as an attractive refuge for smaller startups that are overburdened by the uncertainty of regulation in places like the US and the EU, she said.

To be sure, the UAE has its own regulations, but the centralized power of the Emirati leadership has made the rules more straightforward and easy to adjust when necessary.

The country has also created several special economic zones, or “sandboxes,” where startups can experiment freely with new technologies, from autonomous vehicles to healthcare to crypto.

September 24, 2024 | New York City | Request Invitation

President Lazarus Chakwera, Malawi and Dr. Monique Nsanzabaganwa, Deputy Chairperson, African Union Commission will join the stage at The Next 3 Billion summit — the premier US convening dedicated to unlocking one of the biggest social and economic opportunities of our time: connecting the unconnected.

California governor Gavin Newsom hinted Tuesday that he’s leaning toward a veto of the state’s controversial AI law, SB 1047. He didn’t come right out and say it, but he said he was concerned about a possible chilling effect. “We dominate this space, and I don’t want to lose that,” he said, speaking at Dreamforce, the annual San Francisco conference/soirée put on by Salesforce.

Last month, I wrote about why Newsom was likely to lean this way. The symbolism of this bill is now bigger than the legislation, which has been somewhat defanged, anyway. And it’s even bigger than California. Democrats need to be seen as pro-technology again, or they’ll lose an important source of fundraising support.

On Tuesday, Newsom signed a handful of bills into law that deal with more narrow issues around AI, including one that gave actors more control over their likenesses. That was a smart move, considering the Screen Actors Guild was a supporter of SB 1047.

Racial biases have been known to creep into artificial intelligence algorithms. Now teachers are bringing it into the classroom as they police students’ use of generative AI tools like ChatGPT to complete homework, according to a new study by children’s safety nonprofit Common Sense Media.

The group found that Black teenagers in the US are about twice as likely as their white and Latino peers to have teachers incorrectly flag their schoolwork as AI-generated. Common Sense Media surveyed 1,045 13- to 18-year-olds and their parents from March 15 through April 20.

“This suggests that software to detect AI, as well as teachers’ use of it, may be exacerbating existing discipline disparities among historically marginalized groups,” said the report, which was released Wednesday. “In the United States, Black students face the highest rate of disciplinary measures in both public and private schools — despite being no more likely to misbehave — which contributes to negative impacts, such as lower academic performance.”

The findings aren’t surprising, Robert Topinka, a senior lecturer in media and cultural studies at Birkbeck, University of London told Semafor in an interview.

Part of the problem is that AI detection software is “wholly unreliable,” he said. It’s trained to flag generic and formulaic phrasing using pattern matching, but can’t distinguish between ChatGPT and spelling and grammar checkers like Grammarly, which means students risk being penalized even for using approved tools.

— Mizy Clifton

• Capitol Hill’s bipartisan outrage at Secret Service: ‘They failed twice.’
• Top US antitrust cop says DOJ will overhaul how it reviews bank deals.
• South Africa urges Elon Musk to invest at ‘home.’