A local news reporter covers city government, schools, courts, business, and community life for a daily or weekly paper, a hyperlocal digital outlet, or a nonprofit newsroom — attending public meetings, filing public-records requests, interviewing officials and residents, and turning that into multiple stories a week on tight deadlines, while also posting to social media and sometimes shooting photos or video.
What the audience will actually acceptFigures are percentages. Share of the public comfortable with each kind of news, across six countries. Data: Simon, Nielsen & Fletcher, Reuters Institute (2025). Chart drawn for this page; the underlying figures are the contributor’s verified sources.
How AI is used today
AI use among journalists is already common but concentrated in support tasks rather than reporting. In a representative 2024 survey of 1,004 UK journalists, 56% used AI professionally at least weekly, heaviest in language processing: transcription and captioning (49% monthly), translation (33%), and grammar and copy-editing (30%). More substantive newsroom uses are smaller but real — 22% used AI monthly for story research, 16% for brainstorming, 16% for generating headlines, 12% for fact-checking assistance, and 10% for first drafts. A qualitative study of early-adopter journalists in the Netherlands and Denmark maps these onto the workflow: aggregating sources and spotting trends during newsgathering; transcription, summarization, and headline drafting during production; light-touch human-checked fact-checking during verification; and social captions, newsletters, and SEO during distribution. Verification is where journalists trust AI least — as one reporter put it, "the responsibility of verifying information relies on the individual journalist."
The data suggests if you build it [an AI news product], do not expect overwhelming demand for it.
Benjamin Toff, inverting “if you build it, they will come” for newsroom leaders at the Poynter/AP Summit on AI, Ethics and Journalism, April 2025
What’s changing
Two things are moving in opposite directions, and the gap is the story. Newsrooms expect to use more AI: 60% of UK journalists say their outlet has already integrated it, and far more expect that to grow than to shrink. Audiences are not moving with them. In a six-country 2025 survey, only 12% of the public were comfortable with news made entirely by AI, rising to 21% with human oversight, against 62% comfortable with news made entirely by a human. A U.S. survey found nearly half of Americans uncomfortable with any AI-produced news, and strong demand for disclosure and written ethics guidelines. Journalists are split the same way — 62% see AI as a large threat against 15% who see a large opportunity, though daily users are far more positive. The five-year outlook: newsroom use keeps expanding, and that gap in public trust, not the technology, is what a local newsroom will spend the most time managing.
What a new entrant needs
A reporter entering local news in 2026 needs AI fluency in the low-risk tasks (transcription, translation, copy-editing, first-pass research) and AI scepticism in the high-risk ones (verification, sourcing, judgment about what is newsworthy). That maps onto Critical Thinking and Judgement: summaries and drafts must be checked against primary sources, because hallucinated or oversimplified output can misrepresent a story and damage a reporter's credibility. It also touches Ethical and Responsible Use — because audiences want disclosure and guidelines, a new reporter should know their outlet's AI policy before using any tool on a story and default to disclosure when in doubt. And do not expect AI to simply free up time: frequent AI users in the UK survey were more, not less, likely to say they still spend too much time on low-level tasks.
Where the evidence disagrees
The clearest tension is direction versus reception — newsroom leaders and working journalists expect AI use to keep growing while the public they serve remains largely unconvinced and wants more guardrails. Journalists disagree among themselves about where the line sits: some early adopters were open to AI-assisted drafting for routine beats like sports and financial reporting, while others in the same newsrooms opposed AI-generated text in any published story. For a local reporter that means there is no industry consensus to defer to. A limitation worth naming rather than papering over: these sources cover UK and multi-country journalist and audience surveys plus a Netherlands/Denmark newsroom study, and none is a report exclusively about local U.S. news.
4 sources, located and verified by the contributor, 1 added in review
Production work is commoditised, so judgment and taste become the differentiator.
Research by Sara Nartayeva
Typical employers Software and technology companies · banks and insurers · retailers · healthcare systems · any large organization with a digital product team
What the job is
A UX/UI designer in a corporate setting designs the interfaces and interactions people use to complete tasks inside a digital product — researching how users actually behave, sketching and prototyping screens and flows (usually in Figma), maintaining a shared design system, and partnering with product managers and engineers to ship.
AI absorbs the mechanical end of this job first — first-draft screens, routine asset variations, basic wireframing. What it cannot do is frame the right research question, or notice that a plausible-looking screen is inaccessible, inconsistent, or off-brand.
How AI is used today
AI has moved from novelty to routine tooling. Generative tools such as Figma Make, Uizard, Framer AI, Google Stitch, UX Pilot, and Flowstep turn a text prompt, rough sketch, or screenshot into an editable wireframe or high-fidelity mockup in minutes rather than hours. Designers also use general-purpose assistants for research synthesis — turning raw interview notes or survey data into themes — plus first-draft interface copy and quick layout variations to react to in a design review. In corporate settings AI is increasingly used for accessibility checks, generating developer-ready specs and code snippets at handoff, and analysing usage data to flag where users are struggling.
What’s changing
The best-supported trend is that AI absorbs the most mechanical, lowest-judgment parts first: first-draft screens, routine asset variations, basic wireframing, and first-pass research write-ups. Nielsen Norman Group's 2026 assessment pushes back on the idea that AI is about to replace designers wholesale, arguing the tools have become good at specific tasks while human direction, curation, and verification remain essential. Labour-market indicators point the same way, with BLS projecting much-faster-than-average growth for UI/UX designers through 2034 and the World Economic Forum listing UI/UX design among the fastest-growing roles by 2030. Confidence is not uniform, though: more speculative commentary — anecdotes about entire junior design teams being replaced, forecasts that AI agents will take over most traditional UX work within a couple of years — is far less verifiable and reads more like prediction or marketing than established fact. The day-to-day task mix is shifting quickly and visibly; claims that the role disappears are much shakier.
What a new entrant needs
Visual and interaction fundamentals still matter, because AI tools produce plausible-looking screens that can be inconsistent, inaccessible, or off-brand, and a junior designer needs the judgment to notice. That is Critical Thinking and Judgement: treat an AI-generated mockup the way you would treat a junior colleague's first draft — useful, but requiring evaluation against usability heuristics, accessibility standards, and the design system before it goes near a real user. Build real research skills, since AI compresses synthesis time but cannot frame the right question or build rapport with a user. Learn to prompt and iterate deliberately rather than accepting first outputs. And get comfortable being the human in the loop who is accountable when an AI-assisted design fails. Because production work is increasingly commoditized, standing out as a junior will depend more on judgment, taste, and the ability to explain why a design decision was made than on raw speed at producing screens.
4 sources, located and verified by the contributor
Figma. (2025). AI report: Perspectives from designers and developers.
Nielsen Norman Group. (2026). Assessment of AI's effect on UX work.
Game Developer
AI writes the prototype. It does not survive contact with debugging and integration.
Research by Katrina Golubchik
Typical employers Self-funded independent studios · publishers running hundred-person teams
What the job is
A game developer designs and builds interactive software: the systems, the mechanics, and the code holding them together. The role runs from solo one-person studios to specialised posts at large houses — gameplay programmer, systems designer, technical artist. What holds it together at any scale is one loop: design a system, implement it, playtest it, then cut or keep it depending on whether it is actually fun. For a solo developer that loop is compressed into one person, which is why AI tooling lands differently here than in a studio with separated disciplines.
What the minority who use AI actually use it forFigures are percentages. Share of generative-AI users in the game industry naming each use. Data: Elderkin & Ball, GDC State of the Game Industry (2026). Chart drawn for this page; the underlying figures are the contributor’s verified sources.
How AI is used today
The honest picture is narrower than the discourse suggests. “Vibe coding” — describing a game in plain language and letting an agent write the implementation — is real and genuinely useful for prototyping, particularly for solo developers without a deep engineering background; Claude Code, Cursor, and Copilot have made it normal in small-scale and hobbyist work. Industry-wide, the data is more modest. GDC’s 2026 survey of more than 2,300 professionals found only 36% of game industry workers use generative AI in their job at all, dropping to 30% at studios specifically. Among those who do, the dominant use is research and brainstorming, not authorship. A peer-reviewed synthesis of ten qualitative studies reaches the same conclusion from another direction: the present value is upstream, in ideation, rather than end-to-end production, with human intervention still needed to integrate anything the model produces.
The skill being displaced is not programming judgement. It is syntax recall — and those are not the same thing.
What’s changing
What the evidence supports is narrower than the popular claim, and more interesting. Efficiency gains are conditional rather than general: they hold during ideation and early prototyping and largely disappear once a project reaches refinement, debugging, and integration with an engine’s pipeline. That matches the survey, where even among AI users, coding assistance ranks behind simple research. The industry’s response has fractured along a visible line. Independent and student developers describe generative AI as lowering the barrier to entry; working professionals are considerably more hostile, with 59% of programmers and 63% of designers calling its impact on the industry negative — up sharply from 18% two years earlier. Larian Studios publicly reversed its use of generative AI for concept art after backlash from players and staff.
What a new entrant needs
Knowing how to prompt a coding agent is not a substitute for understanding what the generated code does, and treating it as one is a fast route to shipping something broken. That is Critical Thinking and Judgement in the AI Readiness framework: not avoiding the tools, but assessing their output, catching inconsistencies, and cross-referencing rather than accepting. A junior developer now needs to read and debug generated code as fluently as they write their own — arguably more so, because the failure mode of generated code is plausible-looking logic that is subtly wrong, which is a different problem from your own mistakes.
Where the evidence disagrees
The split is not really about whether the technology works. It is about who carries the risk when it is used carelessly, and that risk falls very differently on a solo hobbyist than on a junior employee whose job is on the line. Worth noting too that the professionals most sceptical of generative AI are concentrated in the disciplines doing the most hands-on technical work — programming, design, visual art — which suggests scepticism and competence are not opposed here but are the same instinct.
3 sources, located and verified by the contributor