A new category of role is appearing that did not exist three years ago. Here is what prompt engineers, AI product managers, and AI ethicists actually do, and whether the roles are durable.
A new category of role has appeared in 2026 that did not meaningfully exist three years ago. These are the AI-native roles, created in response to the fact that organisations now need specific people whose job is to deploy, orchestrate, govern, or build around generative AI systems. The three most visible are prompt engineers, AI product managers, and AI ethicists. Understanding what each one actually does, and whether the role is durable, matters for anyone considering a move into this category.
Prompt engineers
Prompt engineers design and optimise the instructions that AI systems execute. The craft involves understanding how different models respond to structure, tone, and context, and building prompt libraries that produce reliable output at scale. The role grew rapidly in 2023 to 2024, peaked in visibility, and is now evolving. The prompt-as-a-standalone-skill framing is compressing fast because AI tooling is getting better at self-prompting and because most teams now expect every practitioner to be prompt-literate.
The durable version of this role is the one where the work is integrated into platform engineering or AI product ownership. The prompt engineer who also owns evaluation, retrieval, and orchestration sits at the authority threshold. The prompt engineer who only writes prompts is becoming a commodity. For the broader technology landscape see AI risk in technology jobs.
AI product managers
AI product managers are the most structurally durable of the three. They sit at the intersection of product strategy, engineering, and AI capability, and they own the decision of what to build with AI, for whom, and how it should behave. The role is appreciating because every organisation is now shipping AI features, and the ones who have dedicated AI product owners ship better outcomes than the ones who bolt AI onto general product management.
What makes AI product management durable is the accountability layer. The AI product manager is the named decision-maker on model choice, user-experience trade-offs, and the boundary between AI automation and human oversight. That is structural authority, not execution work. For how the technology category is shifting broadly see the technology hub.
AI ethicists
AI ethicists are the youngest of the three roles and the most uncertain in structural outcome. The role covers policy, governance, bias and safety evaluation, and the interface with regulators. In principle it is the most insulated of the three because the work is accountability work. In practice, the role is often positioned as advisory rather than decision-making, which is a weaker structural position.
The durable form of AI ethics work is the one embedded inside product and engineering organisations with real sign-off authority on model deployment. Advisory-only AI ethicists whose recommendations get filed away are in a compressing position even though the topic is growing. The question to ask of any AI ethics role is whether it comes with enforcement scope or just commentary scope.
How to assess an AI-native role before joining
AI-native roles carry novelty premium in compensation today, but novelty is not durability. The structural questions to ask are the same as for any role. Does the work sit above the authority threshold? Does it accumulate judgment, accountability, or scope? Is the skill set tied to a specific model generation that may compress as models mature? The AI Career Index produces these readings structurally so you can evaluate an AI-native offer on durability rather than buzzword weight. For a deeper view on adjacent technology roles see AI engineers.
Run your structural assessment
The AI Career Index is free to take. Get your structural snapshot in 10 minutes, including your durability score, archetype, and projection.
Take the Free Assessment