September 18, 2026

Good morning 👋

Welcome back to OneClick.

AI is everywhere in today's tech news — but the interesting part isn't another model announcement.

It's what AI is starting to change around itself.

Let's get into it. 👀

Amazon has now entered the growing debate over how quickly AI should develop.

The company says AI models should only be released when they are ready and safe to use, with rigorous testing and strong safeguards.

Amazon isn't calling for an industry-wide slowdown. Instead, it says technology companies and governments need to work together on protections as AI advances.

The AI debate is moving beyond:

“What can AI do?”

It's increasingly becoming:

“How do we make sure what AI can do is safe enough to use?”

The UK's King Charles III brought senior figures from major AI companies together in Scotland this week, including representatives from OpenAI, Anthropic, Google DeepMind and Nvidia.

The meeting focused on AI safety and the need for technology to remain under meaningful human control. It comes as the industry debates how quickly increasingly capable AI systems should be developed.

AI safety is no longer just a Silicon Valley conversation.

Governments, companies and society are all starting to wrestle with the same question: what happens if the technology moves faster than our ability to manage it?

Here's a development that deserves a second look.

Anthropic says its Claude AI system is now leading 26% of the company's model research and development work, meaning it can complete substantial tasks from a high-level prompt while remaining under human supervision.

The company says Claude is involved in around 90% of its R&D work in some form.

That doesn't mean Claude is independently building the next model from beginning to end.

But the proportion of work AI is doing is growing rapidly.

We're moving from:

Humans build AI

towards:

Humans and AI increasingly build AI together.

That's a change worth watching.

Huawei has unveiled new AI computing technology in China, including its Atlas 960 SuperPoD system.

The development is another sign that China is pushing hard to build its own AI computing infrastructure as U.S. restrictions limit its access to some of the world's most advanced chips and chipmaking equipment.

The AI race therefore isn't only about who has the smartest model.

It's also about who can build the hardware needed to run those models.

AI needs brains — but those brains need hardware.

The race for AI is also a race for chips, computing power and the ability to make them.

A new report says security researchers used Anthropic's Claude during an authorised bug-hunting exercise to gain access to parts of OpenAI's systems.

The researchers reported the vulnerability to OpenAI and received a $6,500 bug bounty. They did not access OpenAI's model weights.

The incident follows other recent reports of AI agents being used to probe systems and automate cyberattacks.

AI can help defend computer systems.

It can also help find ways into them.

That makes AI capability itself part of the cybersecurity equation.

Apple's latest software rollout includes major updates to its AI capabilities, alongside its new iPhone, Apple Watch and AirPods lineup.

Apple's own newsroom says its latest software updates are now available, including its new, more capable Siri experience and new child-safety features.

The bigger shift is that AI is increasingly becoming something people interact with through ordinary devices rather than a separate chatbot or website.

AI is moving out of the “tech demo” category.

It's becoming part of the devices people already use every day.

Put all of today's stories together and something interesting appears.

AI isn't staying inside the AI industry.

It's reaching into:

🤖 How software is developed

💾 Which chips and hardware countries build

🔐 How cybersecurity works

🇬🇧 How governments and institutions respond

📱 How ordinary devices work

⚡ And, as we saw yesterday, how much electricity technology needs

The question for the next few months isn't only which AI model is smartest.

Watch what happens around the technology:

Who builds the chips?

Who supplies the electricity?

Who sets the safety rules?

Who controls the data?

And how much of the work starts being done by AI itself?

If you're rushing out the door, remember these five:

1. Amazon has joined the growing AI safety conversation.

2. AI safety is now being discussed at the highest levels in the UK.

3. Anthropic says Claude is already leading a significant share of its own AI research work.

4. China is pushing harder to build its own AI chip infrastructure.

5. AI is becoming both a cybersecurity tool and a cybersecurity risk.

AI isn't just changing software anymore.

It's starting to reshape the infrastructure, security, hardware and everyday technology around us.

And that's the part worth watching.

See you next time. 👋

OneClick Tech

The tech news that matters, without the noise.