Tech Workers Adopted AI First. Here Is What Every Other Industry Can Learn From That.
New survey data maps exactly how fast AI is spreading across healthcare, education, finance and manufacturing, and pinpoints the one skill most workers are ignoring.
Key points
- Only 18% of healthcare workers have been asked to use an AI chatbot at work, compared with the majority of tech workers, according to a 2026 TripleTen survey.
- 79% of workers across all industries report being asked to learn at least one entirely new skill because of AI.
- When an employee taught their manager a new AI skill or process, 88% of those employee-led innovations became standard company practice.
- Healthcare workers spent an average of $990 out of pocket on professional development last year; nearly half spent nothing at all.
- Only 30% of managers and owners rank the ability to spot AI errors as an absolute must-have skill, putting it sixth on a list of seven.
Tech workers aren't a special case. They're just first in line.
That's the plainest reading of new survey data published by the education company TripleTen, first highlighted by MIT Technology Review. The numbers show a clear adoption curve: finance is close behind tech, with 48% of finance workers already asked to use AI chatbots, the software tools that generate text and answer questions, and 32% asked to automate parts of their workflow. Healthcare sits at 18%. Education sits at 25%.
Those gaps aren't failures. They reflect the pace at which different industries can safely absorb new tools. A hospital can't roll out an AI writing assistant the way a software firm can. Clinical compliance, patient safety concerns and legal obligations all slow things down deliberately.
What does this mean for workers outside tech?
The window between now and a broad healthcare or education rollout is a genuine advantage, but only if workers use it to build the right skills.
Across every sector surveyed, workers are already being asked to learn new things. The specific skills now common in tech, checking whether an AI output is correct, writing clear instructions for AI tools (called prompts) and automating repetitive tasks, are the same ones likely to arrive in hospitals and schools within a few years.
One skill stands out for being undervalued. Judging whether AI output is right or wrong ranked sixth out of seven skills that managers called essential. Only 30% called it an absolute must. That number should be higher. An AI tool that summarises a patient's records or drafts a student's learning plan can be confidently wrong, and the person who catches the mistake needs enough real-world experience to recognise it. You can't spot a wrong answer if you don't know what the right one looks like.
Our coverage of healthcare AI goes back to 6 August 2026, and the pattern we've seen across six stories holds here: institutional caution is real, but it doesn't protect individual workers from being caught unprepared when rollout finally happens. In our story on how Kier Group trained 11,000 employees, published 2 September 2026, we found that peer-led learning got 40% of staff using AI tools daily, which suggests the employee-to-manager influence the TripleTen data describes isn't unique to tech.
| Sector | Asked to use AI chatbots | Avg. Out-of-pocket training spend | Share who spent $0 |
|---|---|---|---|
| Tech | Majority | $1,758 | 13% |
| Finance | 48% | $1,664 | 24% |
| Education | 25% | $883 | 42% |
| Healthcare | 18% | $990 | 45% |
The training spend figures reinforce this. Tech workers average $1,758 a year on their own professional development; education workers average $883, with 42% spending nothing. Workers who wait for their employer to fund everything tend to move slower than those who invest a little of their own time and money early. Finance workers are close to tech in both adoption and personal investment. Healthcare sits awkwardly in the middle: more cautious than finance on adoption, but spending more per person than education workers while still leaving 45% at zero.
Common questions
If AI is barely used in healthcare right now, should healthcare workers bother learning it?
Yes, and the timing is actually good. The relatively low adoption rate today gives healthcare workers time to learn at a reasonable pace, before AI use becomes a basic job expectation rather than a career advantage.
What is the single most useful thing a non-tech worker can do right now?
Practise spotting AI mistakes in low-stakes situations. That skill, knowing when an AI answer is wrong, transfers across every industry and is the one managers are most likely to underestimate until something goes wrong.
If an employee teaches their manager a new AI skill, does it actually go anywhere?
According to the TripleTen data, yes. In 88% of cases where a manager learned an AI skill from an employee, that practice became standard company policy. Early adopters at the team level can genuinely shape how their whole organisation works.
