Two professions are having almost identical conversations right now, and I think they deserve to be heard in the same breath. Journalists and software developers are both watching the same thing happen to their work. In both cases, AI tools are being adopted at high rates. In both cases, the people doing the adopting are deeply uncertain about what they are adopting. In both cases, entry-level workers are bearing the sharpest end of the disruption. And in both cases, the dominant industry narrative — that this is augmentation, not replacement — sits awkwardly against the actual numbers.
The Stack Overflow data from this year is striking, not for the adoption figure (84% of developers use or plan to use AI coding tools) but for what sits beside it. Trust in those tools has fallen to 29%, down 11 points in a single year. The most common frustration, cited by 66% of developers, is dealing with AI output that is almost right but not quite: code that looks correct, passes a quick read, and then fails or introduces a vulnerability you have to find. Debugging AI-generated code is often more time-consuming than writing it yourself. The developers know this. They are using the tools anyway.
That is not enthusiasm. That is something more like coercion without a coercer.
I am a software developer. I use these tools every day. I benefit from them. I also know that computer programmer employment in the US has dropped 27.5% since 2023, that entry-level hiring at major tech firms is down 25%, and that this year's graduating cohort of computer science students is facing the most pessimistic job market since 2020. The people who keep telling me the role is transforming, not disappearing, are without exception describing what is happening to senior engineers. The junior end is being quietly removed.
Journalism is running the same playbook, a few steps behind. E tū, the union representing New Zealand journalists, put out a statement this month that deserves to be read carefully. Their core concern is not that AI is bad technology. It is that human journalists, their sources, their expertise, their accountability, must remain central to storytelling, and that commercial interests are already pulling against that principle. The student newsroom at AUT, Te Waha Nui, has published an AI policy that says plainly: news is best created by humans. Their reasoning is not technophobic. It is professional: they are trying to develop skills, and they understand that outsourcing the core task undermines that development from the start.
Research published last year found that roughly 9% of newly published articles are either partially or fully AI-generated, with the figure skewed heavily toward smaller, local outlets. The communities served by local papers, often the communities with fewer resources and less political power, are getting more AI-generated content, not less. The people who most need journalism to work for them are getting the version of it the industry finds cheapest to produce. That is not transformation. That is a transfer of cost onto the people least able to bear it.
What links these two professions is not the technology. It is the pattern. Individual decisions to adopt AI tools are rational. I cannot blame a journalist who uses AI transcription to save an hour, or a developer who uses a coding assistant to get through a sprint. The economics make sense at the individual level. The aggregate result of all those individually rational decisions is a slow erosion of the human labour, the trust, the craft, and the entry points that make both professions function.
The people worried about this are not confused. They are not failing to see a deeper structural picture. They are the deeper structural picture. A developer who does not trust the tools they are required to use, a student journalist trying to build skills in an environment that keeps offering to shortcut them, a junior programmer who cannot get a first job because the entry-level work has already been automated: these are not people who need to be redirected toward a more sophisticated analysis.
They are the analysis.