LINGUAWIRE
Google is testing AI chips on a small satellite. In South-East Asia, experts say AI could make false political content faster and more targeted.
A current English news story adapted for language learners at B1, with audio, transcript, vocabulary and comprehension work.
Welcome to LinguaWire.
Today, Google tests AI technology in space, and we examine how AI could increase election disinformation risks in South-East Asia.
If you would like to read along, the transcript and every word explained are waiting at linguawire.com.
Our first story is about Google’s Project Suncatcher and its plans for AI data centers in space.
Sophie is LinguaWire's correspondent for technology.
Thanks, Daniel. Google has launched a refrigerator-sized satellite as part of an experiment called Project Suncatcher.
This is not yet a data center in orbit. It is a small test of whether AI equipment can work safely in space.
That’s right. The satellite carries four of Google’s Tensor Processing Units, or TPUs, which are chips designed for machine learning.
What will those chips do during the mission?
They will run a version of Google’s Gemma AI model for 15 minutes at a time and answer simple questions. The short sessions are necessary because of heat management limits.
So the test is about more than whether the AI model gives useful answers.
Yes. Google wants to see how the chips deal with the physical stress of launch, radiation, and extreme temperatures in space.
Why does Google want to put this kind of computing equipment into orbit?
AI services need a great deal of computing power, and that has increased demand for data centers on Earth. Google believes space could offer a nearly constant supply of solar energy.
The satellite will use a sun-synchronous orbit, meaning its solar panels will almost never be in shadow.
That could reduce the need for heavy batteries or backup power. Google developed the satellite with Planet, and its longer-term idea is to build groups of satellites that work together as space-based data centers.
It shows how companies are looking for new sources of energy as AI infrastructure grows. Thank you, Sophie.
Our second story is about the risk that AI could make election disinformation more powerful in South-East Asia.
Liam is LinguaWire's correspondent for technology.
Good to speak with you, Daniel. Experts are concerned that AI can make false political content faster, cheaper, and more personal.
One recent case in Malaysia has brought that concern into sharper focus.
Anthropic said its Claude AI was used in an operation that appeared designed to influence voters during recent Malaysian state elections.
What did this operation involve?
It used about 1,000 accounts on X, a false news outlet called Malaysia Pulse, and fabricated documents. It also used census and electoral data to target sensitive issues involving race, religion, and royalty.
The operation covered all 222 Malaysian parliamentary constituencies, so this was an attempt to reach people in a very detailed way.
Yes, although Anthropic stepped in and the experiment was ultimately ineffective. Still, the case showed how AI can help create and organise a large influence campaign with less effort than before.
AI may produce the content, but people still need to see it somewhere.
Exactly. Katie Harbath, a former Meta public policy leader for global elections, says social media platforms remain the main way this material is distributed.
Why are experts especially worried about South-East Asia?
The region already has serious problems with fake news, and many people are young and highly connected online. Maria Ressa, the Filipino journalist and Nobel laureate, has said techniques used in the Philippines were later copied in western countries.
So AI does not create disinformation from nothing, but it can make an existing problem easier to scale. Thank you, Liam.
Part one: vocabulary.
Influence.
To influence means to affect what someone thinks or does; in the story, “an operation that appeared designed to influence voters”.
Social media can influence public opinion.
Target.
To target means to direct something at a particular person or group; in the story, “to target sensitive issues involving race, religion, and royalty”.
The campaign targets young voters.
Scale.
To scale means to increase something so it reaches more people; in the story, “it can make an existing problem easier to scale”.
The company wants to scale its online service.
Part two: the grammar.
Listen to this: “AI may produce the content, but people still need to see it somewhere.”
Here, it is an object pronoun: it replaces the content after the verb see.
Again, “AI may produce the content, but people still need to see it somewhere.”
I read the report and shared it with my team.
Repeat after me: AI may produce the content, but people still need to see it somewhere.
English news for learners · English podcast · English Coursebook · English level test · Today's episodes · How it works · Podcast and video · Programme · News · Coursebook · Level check · For students and exams · Weekly crossword · Enrol