Almost half of South Africa's AI practitioners say AI has made them feel more secure in their jobs. In DayOne's State of AI Delivery ZA 26 report, 47% of the 282 people surveyed said AI had made them feel more secure in their role over the past year, compared with 24% who felt less secure. The other 29% said nothing had changed.
Overall, 51% said they are more excited than concerned about AI's impact on their field.
Insights on AI job security
How people feel about AI appears to depend a lot on how they formed their expectations of it. Most respondents, 47%, said their expectations came from their own research and testing. Another 24% pointed to media and industry hype, 15% to their peers, 12% to a leadership mandate at work and 2% to claims by vendors or consultants. Among those who formed their views through their own research, 54% feel more secure, and 57% are more excited than concerned.
The group whose expectations came from a leadership mandate looks very different. It’s the only group in which more people feel less secure than more secure: 42% less secure against 27% more secure. Only 27% of that group are more excited than concerned, while 48% say they are evenly split. That group is small, about 34 people, and the report notes that these comparisons show links, not causes.
Leadership plays a big role in adoption more broadly: 48% of respondents said AI adoption at their company is driven by a leadership mandate, against 45% who said it is led by staff from the ground up, and at companies with more than 200 people, 64% said leadership drives it.
Asked what excites them most, respondents most often named speed, getting more done in less time. "AI lets me build working prototypes and small applications myself, end to end, without waiting for development capacity," one said.
The respondents were surveyed between 2 and 16 October; 62% are in technical roles, and 29% are self-employed or freelance, so for some, "secure in my role" means their own business rather than a job.
What others are saying about AI job security
DayOne, an AI build studio and network of AI professionals, sums up its wider finding as "adoption is solved, delivery isn't", pointing to people and resources rather than technology as the main gap. In Monday's look at small business AI, we covered what the report found about how teams set their expectations. Our recent piece on AI training budgets found half of these companies have no budget to train their staff.
Being told is not the same as finding out
The headline here is reassuring. Among people who use AI closely, twice as many feel it has made their job more secure than less, which runs against the common fear that AI mainly threatens jobs. The more useful finding for employers is the difference between groups.
People who tested AI for themselves tend to feel more secure and more excited. People whose expectations were set by a mandate from above are the one group where the feeling tips the other way. The data cannot say a mandate causes that, and the group is small. But it matches what you would expect: being told to use a tool, without the time to understand it, can feel like a threat rather than a help.
For managers rolling out AI, the practical lesson is to give people room to test and learn before setting targets, which costs little and, as our training budget piece found, is often missing. Disclosure: DayOne's survey was promoted in The Open Letter.
You might also like our piece on Claude vs ChatGPT in SA, women's ChatGPT use in SA, and SA's AI policy.
Get more SA tech and business news and subscribe to The Open Letter.


