What people are actually worried about with AI.

Since May 2026 I've spoken to over 30 people and had more than 100 conversations about AI, from big companies to small businesses. After each one I wrote down what I'd heard. These are the worries that came up most. (Just the worries. The optimists can have another article.)

What are businesses most worried about with AI?

Rank	Worry	Organisations that raised it
1	Where does our data go?	7
2	Will it take my job?	6
3	Nobody has told us the rules	5
4	Can we trust the answer?	5
5	Is it worth the money?	5
6	It looks AI-made	3
	The AI isn't good enough	0

Nobody, not once, said the AI isn't good enough. Data came up first, often in the very first meeting.

Then, while writing this up, I noticed something. The top worries all came true in two weeks of AI news, in late October 2026. Everything below is as of 30 October 2026.

Worry 5: is AI worth the money?

On 29 October 2026 (US time), Google's parent company lost nearly 13% of its value, about US$500 billion, after its quarterly results. Its AI business is booming, with Google Cloud up 103% in a year. What spooked investors was the bill: around US$200 billion this year, mostly on AI, and more to come.

It isn't only Google. China's STAR 50, an index of big Chinese tech and chip companies, was down about 40% in under four months.

So is it a bubble? Partly, I think. In 2000 the internet was real, and plenty of money was still wasted. AI is the same: real value, not another Web3 hype cycle, but nobody has proven who makes the money back. The real test: that same week, Anthropic, the company behind Claude, filed to list on the US stock market.

What it means for you: nobody needs to bet like Google. Pick one job where you can see AI paying for itself, and prove it there first.

Worry 2: will AI take my job?

This is the interesting one. In just three weeks:

Company	What happened
Microsoft	Offered exit packages to about 60,000 staff, a quarter of its workforce, to cut costs and spend more on AI.
Barclays	5,000 jobs. The bank said it outright: "a result of increased investment in AI".
Amazon	250 roles, mostly developers, a day after a memo telling engineers to embrace AI.
Westpac	140 more roles, after 102 two weeks earlier. The union blames AI. The bank says priorities.

My read: this is the AI pivot happening. Companies are betting AI lets fewer people do the same work, and moving money from salaries to computers before it's proven.

Some will get it wrong. Just under 1 in 3 managers who cut a role mainly because of AI later hired for it again (Robert Half, US and Canada). What it means for you: cut after AI is doing the work, not before.

As for me, I'm doing everything I can to stay AI native. I might still end up redundant, but at least I'll know exactly why.

Worry 1: where does our data go?

Four launches in one fortnight, and one idea for your business from each.

Launch	What happened	What it means	Try this
Gemini 4 Carbon, Google, 28 Oct, not out yet	Google's new top model. It's so good at finding holes in software that cyber defenders get it first.	AI can find the weak spots in your systems faster than most IT teams. So can attackers.	Ask whoever looks after your IT if they use AI to check your systems yet.
GPT-6.1 Terra, OpenAI, 26 Oct, paid ChatGPT	Faster, half the price, and it uses a computer like you do. Hacking rating: "Critical", the top level.	AI is moving from writing about screen work to doing it.	List the tasks where someone copies from one screen into another. Those go first.
Genie 4, Google, 27 Oct, testers only	Describe a scene and it builds a world you can steer, with sound and characters who talk.	The gap between watching a video and being inside one is closing.	Practise the hard moments safely: an angry customer, a site emergency, someone's first day.
Claude Duet, Anthropic, 21 Oct, US waitlist	An assistant that checks your apps and messages you first. It asks before it sends, pays, signs or deletes.	AI is moving from something you ask to something that asks you, following OpenAI's Dots.	Before you switch one on, write down what it may never do without a person.

Notice the pattern? Two of the four came with warnings that they're very good at hacking. Worry 1 isn't going anywhere.

What should a business do about these worries?

Start with the one your people raise first, which is usually data. Write down where your information is allowed to go before anyone switches a tool on. Then pick one job where AI can pay for itself, prove it there, and only change roles once the AI is actually doing the work.

If you want help working out where AI fits into how your business already works, book 25 minutes.

Common questions

What is the biggest worry businesses have about AI? Where their data goes. Seven organisations raised it, more than any other worry, and it often came up in the very first meeting.

Is AI a bubble? Partly, in my view. Like the internet in 2000, the value is real and plenty of money will still be wasted, because nobody has yet proven who makes the money back. That is my view, not financial advice.

Are companies cutting jobs because of AI? Some say so outright: Barclays cut 5,000 jobs as "a result of increased investment in AI". But just under 1 in 3 managers who cut a role mainly because of AI later hired for it again (Robert Half).

Did anyone say AI isn't good enough? No. In over 100 conversations, nobody raised the quality of the AI itself as a worry.

---

The worry league is my notes on work conversations, counted by organisation: small numbers, one person's network, no names. News sources: CNBC, IBTimes, CTV, InformationWeek, ACS, Google, Cybersecurity News and Anthropic. My view, not financial advice.

This started life as Build Notes Nº 10, in October 2026.