close
close

AI agents and how to detect a lie

Plus: Chinese electric vehicles have encountered an EU-shaped blockade

This is today's edition of The Download, Our weekly newsletter offers you a daily dose of news from the world of technology.

What are AI agents?

When ChatGPT was first released, everyone in the AI ​​industry was talking about the new generation of AI assistants. But last year, that excitement turned to a new target: AI agents.

Agents played a major role at Google's annual I/O conference in May, when the company unveiled its new AI agent called Astra, which users can interact with via audio and video. OpenAI's new GPT-4o model was also referred to as an AI agent.

This story is only available to subscribers.

Don't settle for half the story.
Get paywall-free access to technology news for here and now.

Subscribe now
Already a subscriber? Sign in

And it's not just hype, although there's definitely some of that. Tech companies are pouring huge sums of money into developing AI agents, and their research efforts could lead to the kind of useful AI we've been dreaming about for decades. Many experts, including Sam Altman, say they're the next big thing. But what are they? And how can we use them? Read the full story.

—Melissa Heikkilä

AI lie detectors can detect lies better than humans

Can you spot a liar? I imagine this question has been on many people's minds lately, given the numerous televised political debates. Research has shown that we are generally pretty bad at distinguishing truth from lies.

Some believe AI could improve our chances and be better than old-fashioned methods like lie detector tests. AI-based lie detection systems could one day help us distinguish facts from fake news, evaluate claims, and possibly even spot lies and exaggerations in job applications. The question is whether we will trust them. And whether we should. Read the full story.

—Jessica Hamzelou

This story is from The Checkup, our weekly health and biotechnology newsletter. Log in to receive it in your inbox every Thursday.

The must-reads

I've scoured the internet to find you the most entertaining/important/scary/fascinating stories about technology today.

1 The EU plans tariffs on electric vehicles made in China
The decision is a massive blow for automakers hoping for a respite from the ongoing trade war. (WSJ $)
+ Chinese authorities are allowed to use Teslas for the first time. (Bloomberg $)
+ The country's electric vehicle battery manufacturers want to enter the stationary energy storage sector. (Reuters)
+ Europe's best-selling Chinese electric car manufacturer has a surprising name. (MIT Technology Review)

2 Xenophobia is widespread in China’s social media
Extreme Chinese nationalism appears to be fueling violent attacks on foreigners. (NYT $)

3 Cloudflare has launched a tool to fend off AI bots
The cloud company's model detects bots that try to steal data from its websites. (TechCrunch)
+ Overall, these are not good times for cloud companies. (FT$)

4 Afghan women lead a secret life on the Internet
They are turning to the Internet to fight against the Taliban's restrictions on their freedom. (WP $)

5 Political candidates track their stolen campaign signs
With a little help from Apple AirTags. (WSJ $)

6 An “ethical” AI music generator cannot create good songs
Professional musicians were unimpressed by the dubious compositions. (Wired $)
+ Training AI music models will soon become very expensive. (MIT Technology Review)

7 map apps are extremely simple
Their need to appeal to a wide audience leaves few feeling satisfied. (The Atlantic $)

8 WhatsApp is experimenting with AI-generated avatars
Good advice: Don't do it. (The Verge)

9 TikTok users are hungry for political content
As the British general election proved. (The Guardian)
+ Three technology trends that will shape the 2024 elections. (MIT Technology Review)

10 Minecraft aims for a future beyond video games
I would bet that AI will probably play a role in these plans. (Bloomberg $)
+ The Facebook gaming giant FarmVille still exists. (The guard)
+ A bot that has watched 70,000 hours of Minecraft could unlock the next big thing in AI. (MIT Technology Review)

quote of the Day

“I'm so sweet. Please watch my campaign broadcast.”

— Airi Uchino, a candidate in the upcoming Tokyo gubernatorial election, is taking a novel approach to win residents’ votes, the Associated Press reports.

The big story

How culture promotes foul play on the Internet and how the new “Upcode” can protect us

August 2023

From Bored Apes and Fancy Bears to Shiba Inu coins, self-replicating viruses and whales, the internet is teeming with scams, hacks and frauds.

And while new technologies come and go, they do little to change the fact that illegal activity exists on the Internet because some people are willing to act illegally and others fall for the stories they are told.

Ultimately, online crime is a human story. But why does it work and how can we protect ourselves from falling victim to such schemes? Read the full story.

—Rebecca Ackerman

We can still have beautiful things

A place of comfort, fun and distraction that will brighten your day. (Any ideas? Write to me or tweet it to me.)

+ These ducks just love the hose.
+ We can't say for sure, but popcorn was probably invented to preserve corn for longer periods of time 🍿
+ Congratulations to Patrick Bertoletti, who won Nathan's annual hot dog eating contest after eating a whopping 58 hot dogs in 10 minutes.
+ Hurry up, George RR Martin, we want another book!