Ice Lounge Media

Ice Lounge Media

The Download: how criminals use AI, and OpenAI’s Chinese data blunder

IceLoungeMedia IceLoungeMedia

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.

Five ways criminals are using AI

Artificial intelligence has brought a big boost in productivity—to the criminal underworld.

Generative AI provides a new, powerful tool kit that allows malicious actors to work far more efficiently and internationally than ever before. Over the past year, cybercriminals have mostly stopped developing their own AI models. Instead, they are opting for tricks with existing tools that work reliably.

That’s because criminals want an easy life and quick gains. For any new technology to be worth the unknown risks associated with adopting it—for example, a higher risk of getting caught—it has to be better and bring higher rewards than what they’re currently using. Melissa Heikkilä, our senior AI reporter, has rounded up five ways criminals are using AI now.

 OpenAI’s latest blunder shows the challenges facing Chinese AI models

Last week’s release of GPT-4o, a new AI “omnimodel”, was supposed to be a big moment for OpenAI. But just days later, it feels as if the company is in big trouble. From the resignation of most of its safety team to Scarlett Johansson’s accusation that it replicated her voice for the model against her consent, it’s now in damage-control mode.

On top of that, the data it used to train GPT-4o’s tokenizer—a tool that helps the model parse and process text more efficiently—is polluted by Chinese spam websites. As a result, the model’s Chinese token library is full of phrases related to pornography and gambling. This could worsen some problems that are common with AI models: hallucinations, poor performance, and misuse. 

But OpenAI is not the only company struggling with this problem: there are some steep challenges associated with training large language models to speak Chinese. Read our story to learn more.

—Zeyi Yang

This story is from China Report, our weekly newsletter giving you the inside track on tech in China. Sign up to receive it in your inbox every Tuesday.

The must-reads

I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.

1 AI just got a little less mysterious
Anthropic has delved into how artificial neural networks work. (NYT $)
+ Understanding more about how AI makes decisions could help us control it.  (Wired $)
+ Large language models can do jaw-dropping things. But nobody knows exactly why. (MIT Technology Review)

2 Google is testing ads in its AI-generated search results
Sponsored query answers? No thanks. (Reuters)
+ Why you shouldn’t trust AI search engines. (MIT Technology Review)

3 China has created a chatbot trained on the thoughts of Xi Jinping
But we’ll have to wait to see how popular that’ll be, as it’s still a way off from being released to the wider public. (FT $)
+ Why the Chinese government is sparing AI from harsh regulations—for now. (MIT Technology Review)

4 Our drinking water is major hacking target🚰
Default passwords are to blame. (IEEE Spectrum)

5 Humane is looking for a buyer
Just weeks after its AI pin device got slated in reviews. (Bloomberg $)

6 How a massive corporation covered up the dangers of forever chemicals
And kept selling them afterwards. (New Yorker $)
+ The race to destroy PFAS, the forever chemicals. (MIT Technology Review)

7 Inside the fight for America’s broadband
Campaign groups are clashing with service providers over access. (Ars Technica)

8 Sailboats are making a comeback
And the sails have had a high-tech makeover. (Economist $)

9 Can beef ever really be climate-friendly?
The US branded a meat packer environmentally friendly. Pressure groups aren’t so sure.  (Undark Magazine)
+ How I learned to stop worrying and love fake meat. (MIT Technology Review)

10 Admire the beauty of Earth from the ISS
These new photographs are truly breathtaking. (The Atlantic $)

Quote of the day

“I wish we had called it ‘different intelligence’. Because I have my intelligence. I don’t need any artificial intelligence.”

—Satya Nadella, Microsoft’s chief executive, is worried about people giving AI systems too much credit, he tells Bloomberg.

The big story

Bringing the lofty ideas of pure math down to earth

April 2023

—Pradeep Niroula

Mathematics has long been presented as a sanctuary from confusion and doubt, a place to go in search of answers. Perhaps part of the mystique comes from the fact that biographies of mathematicians often paint them as otherworldly savants.

As a graduate student in physics, I have seen the work that goes into conducting delicate experiments, but the daily grind of mathematical discovery is a ritual altogether foreign to me. And this feeling is only reinforced by popular books on math, which often take the tone of a pastor dispensing sermons to the faithful.

Luckily, there are ways to bring it back down to earth. Popular math books seek a fresher take on these old ideas, be it through baking recipes or hot-button political issues. My verdict: Why not? It’s worth a shot. Read the full story.

We can still have nice things

A place for comfort, fun and distraction to brighten up your day. (Got any ideas? Drop me a line or tweet ’em at me.)

+ Spare a thought for the Kyles of the world, 706 of whom traveled to the city of Kyle, Texas, only to be told they hadn’t broken a world record.
+ Why are spirographs so hypnotic?
+ If you’re into Impressionism, there’s a whole host of impressive-looking shows taking place this year.
+ Here’s what went down when the Beach Boys met the Beatles.