He can make AI do anything. He has built his identity around a skill that may not exist in three years.
Literal meaning: An AI-ninja is someone who positions themselves, or is positioned by others, as exceptionally skilled at using AI tools: getting the best outputs, automating complex workflows, unlocking capabilities others cannot reach. The ninja metaphor signals mastery, speed, and a kind of technical mystique.
Origin: The term emerged from tech and startup culture around 2022, in job postings, LinkedIn profiles and workshop descriptions, alongside the mainstreaming of large language models. It belongs to a family — rockstar developer, 10x engineer, guru — and that family has been measured. Gaucher, Friesen and Kay (2011) showed across five studies that masculine-coded wording in job advertisements makes women feel they do not belong, without changing what they think they can do; Textio’s analysis of fifty million postings put ninja and rockstar among the words after which women stop applying (Fortune, 2016). The ninja is a recruitment word with a known effect.
A professional identity built around AI tool proficiency, framed as rare expertise, dependent on conditions that do not last.
The Appeal: In a labour market disrupted by AI, positioning yourself as someone who masters AI rather than someone displaced by it is a rational strategy. The identity offers status, relevance, and income. For individuals entering or navigating the workforce, it is a real and often effective move.
The Friction: The identity is built on a dependency. Prompt Engineer — the related profession of directing AI through language — faces the same structural instability: as models improve, the skill required to use them shrinks. That is not a forecast. Zhou et al. (2023) had a model write and select its own instructions and matched or beat human-written prompts on 19 of 24 tasks; Khan (2025), on one benchmark, found that a rule-heavy prompting method lifted GPT-4o and lowered GPT-5: a guardrail turned into a handcuff. Deskilling — automation making foundational skills redundant — is not only what AI does to other professions. Crowston and Bolici (2025) describe when working with AI builds skill and when it erodes it; relying on output without understanding the system is the eroding case. AI Literacy is the more durable counter-frame, and the VU Amsterdam student handbook of 2025 shows what it has come to mean: how the models work, what they cost, what they do to power, with prompting as one item among these. The vocabulary points the other way too. Broligarchy — tech power organised through male bonding — is where the rockstar-and-ninja register ends up at the top; the job ad is where it starts.
Why This Matters: The AI-ninja identity belongs to a moment in technological transition, when tools are powerful enough to impress but complex enough to require specialisation. That window closes. What you built during it either transferred into deeper understanding or it did not, and Lee, Sarkar et al. (2025) measured which way that tends to go: among 319 knowledge workers, the more they trusted the AI, the less critical thinking they did. The term is four years old and there is no research on the identity itself yet. What has been measured is the mechanism beneath it.
Related terms: Prompt Engineer · AI Literacy · AI Dependency · Deskilling · Cognitive Offloading · Hustle Culture · Broligarchy
Read more:
Primary:
- Zhou et al. (2023) — Large Language Models Are Human-Level Prompt Engineers. ICLR 2023
- Gaucher, Friesen & Kay (2011) — Evidence That Gendered Wording in Job Advertisements Exists and Sustains Gender Inequality. Journal of Personality and Social Psychology, 101(1), 109–128
Secondary:
- Crowston & Bolici (2025) · Lee, Sarkar et al. (2025) · Khan (2025) · Genkina (2024) · Fortune (2016) · VU Amsterdam (2025) · Gender Decoder (n.d.) — research, journalism, an institutional handbook, and a tool built on the primary source