Meet the AI Granny that wastes the time of scammers

Welcome to this edition of Loop!

To kick off your week, we’ve rounded-up the most important technology and AI updates that you should know about.

‏‏‎ ‎ HIGHLIGHTS ‏‏‎ ‎

• Why 40% of AI data centres may face power shortages

• Meta’s plans to test VR in education and build “metaversities”

• How Atlas are trying to re-imagine geospatial software

• … and much more

Let's jump in!

1. TSMC awarded $11.6 billion in CHIPS funding by US Government

President Biden has issued the final CHIPS award for TSMC, which is a leading manufacturer of advanced computer chips.

The final agreement with TSMC will also result in $65 billion of private investment.

This money will be used to build three state-of-the-art facilities in Arizona and create tens of thousands of jobs by the end of the decade.

It’s planned that one of those factories will open in early 2025, which will allow the US to build advanced semiconductors.

For many years, that has been outsourced to other nations - primarily Taiwan.

But with the huge demand for these chips, which are now critical for the US economy, President Biden has sought to reverse this and bring the manufacturing jobs back to America.

2. Meta plans to test VR in education, will create digital twin 'metaversities'

The company is partnering with 13 universities to test how virtual reality (VR) could be used within education.

These universities will test Meta’s VR prototypes in their classrooms, before the technology is officially launched.

It’s also bringing its "digital twin" program to Europe, which is starting with a handful of universities in the UK, Spain, and Germany.

Two years ago, Meta adopted a similar strategy in the US and they were able to replicate campuses in the virtual world.

A digital twin is essentially a virtual copy of the real-world. Many factories have been exploring how they can use the technology to spot issues with machinery and improve overall efficiency.

But Meta has other motivations here. For several years, they have sold AR glasses that have cameras and display information as you look around.

These new programs will build on this work, as they hope to spot use cases for the technology - specifically in education - and encourage millions of students to buy the glasses.

3. This AI Granny is happy to talk with phone scammers all day

In the UK, one of the largest phone operators has unveiled an “AI Granny” to waste the time of fraudsters.

It’s a really innovative way to apply some of the latest advances, as these scammers will be kept away from real customers.

The AI voice is indistinguishable from a real person and can interact with scammers in real-time, without the need for a human to guide the responses.

In some cases, she has kept scammers on the phone for over 40 minutes and talked about her love of knitting.

Cyber criminals have stolen £1.17 billion ($1.5 billion) in last year, with many of them exploiting the elderly to do this.

It’s really encouraging to see the technology being used in such an innovative way, especially against those who target the most vulnerable in society.

4. OpenAI plans to launch an AI agent in 2025

For some time now, leading AI companies have talked about developing agents that can control your computer and autonomously solve tasks - rather than forcing you to do it yourself.

According to Bloomberg, OpenAI is now preparing to release their own autonomous AI agent in January.

It follows similar moves from their competitors, with Anthropic recently introducing its "computer use" capability and Google preparing its own version for December.

AI companies, like OpenAI and Anthropic, are under huge pressure to monetise their costly models.

The cost of developing new models is going through the roof, but they can’t keep relying on investors to fund this work.

Instead, they hope that autonomous agents will be their next breakthrough product and will quickly boost revenues.

5. Nearly half of AI data centres may not have enough power

A recent report from Gartner has revealed that AI's insatiable appetite for electricity is expected to surge in the coming years, potentially leading to power shortages for data centres.

In 2022, new servers demanded 195 terawatt-hours of electricity - which is equivalent to the consumption of 18 million households in a year.

However, by 2027, new servers could require a staggering 500 terawatt-hours - or the equivalent of 46 million households.

If we don’t use other sources for net-zero power, the pollution generated by AI data centres could skyrocket.

Microsoft, Meta, and Google are pinning their hopes on nuclear power to avoid the looming crisis, but regulators have blocked some of those plans.

Climate Trace releases data about world’s top polluting cities

At the UN Climate Change Conference (COP 29), Climate Trace and Al Gore released their latest data about global emissions.

Their platform now includes emissions data for over 660 million assets.

Climate Trace is an independent monitoring group and includes dozens of organisations, with each specialising in a specific sector - such as agriculture, marine life, or power generation.

It’s really impressive what they have been able to achieve and I first came across the group in 2022, when their co-founder was speaking in Toronto.

Their data is used to monitor changes in emissions over time, with incredibly detailed estimates - including the pollution generated by individual factories and power plants.

These insights can then be used to encourage companies and politicians to take action against their biggest polluters.

To do this, Climate Trace uses a combination of satellite data and machine learning to estimate emissions.

In some cases, they can use this data to calculate how much money a power plant is making - which is useful information for investors.

If you want to find out more about the organisation, or look at their latest data, you can click the link below.

Google releases a new tool for students

When compared to the likes of ChatGPT, “Learn About” is pretty different. The name’s not great, but coming up with good names is tricky.

This tool is pretty interesting. Instead of offering simple facts like ChatGPT would, their tool tries to be more interactive and is focused on helping you learn.

Some of the answers look very similar to something you’d read in a textbook, with “why it matters” explainers and short activities to test your knowledge.

For example, if you ask “How big is the universe?”, the tool will give you a short answer and includes resources from other websites - such as YouTube - to learn more.

In contrast, ChatGPT will just try to give you the answer - which often includes lots of text and might be incorrect.

Tools like these can be really useful for students, but I’m sure it’s also a godsend for the parents and teachers that have to answer a dozen questions per day…

🎵 Spotify now has over 252 million subscribers

🤖 X sues California over an election deepfake law

🚕 Zoox rolls outs its robotaxi in San Francisco and Las Vegas

🎥 Apple is planning to release a security camera in 2026

🧬 23andMe cuts 40% of staff, following their recent woes

🩸 Wing uses drones to deliver blood samples in London

🔍 Hackers have breached US telecom giants to steal wiretap data

🚗 Pony AI, a Chinese autonomous driving startup, aims for a $224 million IPO

☁️ FTC is now investigating Microsoft's cloud business

🔷 Millions of people join Bluesky as a Twitter alternative

🚔 FBI raids home of Polymarket CEO Shayne Coplan

🌱 Microsoft hopes carbon removal will offset its growing AI emissions

Atlas

Atlas is working to make geospatial data more accessible and user-friendly. Essentially, they’re trying to build a Figma-like tool for satellite imagery.

You can use their platform to view satellite data and add your own data on top.

The company was founded in 2021 by a group of students, after they realised that the industry was relying on expensive and old software.

There are a lot of industries that could benefit from this - from construction to renewable energy, or even urban planning.

It’s also useful for journalists, as they have become reliant on satellite imagery in recent years - especially when it comes to reporting the war in Ukraine.

Following the write-up on Climate Trace earlier, it seems like a good tool if you want to explore some of their data.

You can use the link below to check it out.

This Week’s Art

We’ve covered quite a bit this week, including:

• The $11.6 billion in funding that was awarded to TSMC

• Meta’s plans to test VR in education and build “metaversities”

• The innovative AI Granny that talks to scammers and wastes their time

• OpenAI’s plans to launch an AI agent in the coming months

• Why 40% of AI data centres may face power shortages

• Climate Trace’s data about the world’s top polluting cities

• Google’s new tool to help students learn

• And how Atlas are trying to re-imagine geospatial software

Have a good week!

Liam

Share with Others

If you found something interesting in this week’s edition, feel free to share this newsletter with your colleagues.

About the Author

Liam McCormick is a Senior AI Engineer and works within Kainos' Innovation team. He identifies business value in emerging technologies, implements them, and then shares these insights with others.