Do AI keywords on your CV need 'super intelligence'?

Did the White House really rename AI, and does it change your CV?

No — nothing on your CV needs to change. In late September 2026, the White House published an executive order titled 'Inaugurating The Era Of Super Intelligence', directing federal departments and agencies to use the terms 'Super Intelligence' and 'SI' in official correspondence, public communications, websites, reports and other non-statutory materials, and to stop acknowledging 'artificial intelligence' and 'AI'. On the same occasion, a group of major technology chief executives signed a separate voluntary accord on frontier SI responsibilities that the president described as 'morally binding' and not legally enforceable. Both things happened. Neither one touches the software that reads your CV, and neither changes a single word in the job ad you're about to answer.

It helps to be precise about what an order like this can and can't reach. The text governs how executive branch agencies write, and it even tasks the president's science and technology adviser with developing a federal definition of SI. What it does not do is reach into a private employer's careers page, rewrite the requirements block of a posting, or retrain the parser inside a recruiting platform. Hiring language moves on its own timetable, driven by what engineering managers call their own stack. Expect job ads to keep saying 'machine learning engineer' and 'generative AI' for a long while, because that's what the teams doing the hiring call the work.

So treat the rebrand as a useful stress test rather than a to-do item. The question it raises — which exact term does the system match? — is the single most important question in keyword strategy, and it has the same answer every time: the term in front of you in the posting. That's also why a generic once-and-done keyword list ages badly. The free CV analysis at HRLens lets you upload your CV and paste the job description so the comparison happens per application, against that employer's wording, instead of against whatever vocabulary made the news this week.

callout_placeholder

How does ATS keyword matching actually pick which AI terms count?

Applicant tracking systems score your CV against terms pulled from the job description itself, and the matching is still largely literal. If the posting says 'React.js' and you wrote 'ReactJS', a strict string match can miss it. If it says 'natural language processing (NLP)' and you only wrote 'NLP', you've matched half of what the employer indexed. Newer platforms layer in some natural-language understanding, so 'led' and 'managed' may score similarly, but you should never bet an application on that generosity. Write the employer's exact phrase at least once, then vary the wording naturally in the rest of your bullets so the document still reads like a person wrote it.

Repetition inside the posting tells you what to prioritise. When a term shows up in the required skills, the preferred skills and the responsibilities, the employer has effectively told you it carries the most weight. Mine two sections hard — requirements and responsibilities — and lift the phrasing verbatim where it's honest. Titles behave the same way. 'Head of Data' and 'Director of Machine Learning' can describe identical work, but a system searching for one won't necessarily surface the other, so mirror the company's title language whenever it genuinely matches what you did. Acronyms deserve both forms: spell it out, then put the short form in brackets.

There's a second match most candidates forget. After your CV is parsed into fields, it sits in a searchable database, and a recruiter later runs keyword searches across that database to build a shortlist — sometimes weeks after you applied, sometimes for a different role entirely. That search is typed by a human with their own habits, which is one more reason to include both 'large language model' and 'LLM', both 'generative AI' and the specific model families you've worked with. Being findable twice, by the scoring pass and by the recruiter's own query, is worth more than any clever phrasing you could invent.

Which AI keywords should you actually put on your CV in 2026?

Put down the concrete nouns: the techniques, the frameworks, the platforms and the model families you have genuinely used. 'Machine learning', 'deep learning', 'natural language processing', 'computer vision', 'large language models (LLMs)', 'retrieval-augmented generation (RAG)', 'prompt engineering', 'fine-tuning', 'MLOps', 'vector database', 'model evaluation'. Then the named tools — Python, PyTorch, TensorFlow, scikit-learn, Hugging Face, LangChain, Azure OpenAI, Amazon SageMaker, Databricks, Vertex AI. Specificity is what separates a credible AI line from a decorative one. A recruiter can't verify 'passionate about AI', but they can absolutely ask you about the retrieval pipeline you built, the embeddings model you chose and why you dropped the first one.

Non-technical candidates need this section too, and they're often the ones who get it wrong. If you're in marketing, operations, finance, HR or law, the honest keywords describe adoption and governance rather than model training: 'AI tooling', 'generative AI workflows', 'prompt libraries', 'AI-assisted drafting', 'automation', 'AI policy', 'data quality', 'human-in-the-loop review', 'vendor evaluation'. Pair each with the platform you actually used and the outcome you actually got. 'Rolled out an AI drafting workflow across a 12-person support team, cutting first-response time by a third' tells a hiring manager far more than any list of technologies, and it still carries the keyword.

One term to skip: 'super intelligence'. No employer's requirements block asks for it, so it earns you nothing in matching, and dropping a fresh political label into your skills line invites an opinion you didn't want in the room. The same logic applies to every other vocabulary fashion — 'agentic AI' and 'AI agents' are worth writing only once teams in your field are writing them in postings, and only if you've shipped something that fits the description. Let the job ads lead. Your CV is a mirror of the role you're chasing, not a position paper on what the technology should be called.

table_placeholder

Where on your CV should AI skills go so a parser reads them?

Give them two homes: one plain skills block near the top and the same terms again inside your experience bullets. Keywords in bullets carry more weight with both software and humans, because they arrive attached to evidence. The skills block exists so a literal match and a recruiter's database search can find the exact string fast. Keep it as a simple comma-separated line or a short list under a conventional heading — 'Skills', 'Technical Skills' — grouped by category if the list is long. No ratings out of five, no star icons, no skill bars. A parser reads none of that, and a hiring manager doesn't believe it.

Formatting is where perfectly good AI keywords quietly disappear. Two-column layouts can be read in the wrong order, so a skills sidebar ends up spliced through your job history. Text inside images, logos or shapes isn't text at all once it's extracted. Tables, headers and footers are inconsistently handled, and clever section names like 'My Toolkit' can leave a parser unsure what it's looking at. Stick to one column, standard headings, a normal font, and a PDF with a real text layer — if you can't select and copy the words in your own file, neither can the system. A free ATS resume check will show you what survives extraction.

Order matters more than most people expect. For technical roles, a short skills block above your experience works because screeners look there first; for career changers, it does double duty by surfacing capabilities your job titles hide. Keep dates in a consistent format, one role per entry, and the employer's own title language where it's accurate. If you're rebuilding from scratch rather than patching an old file, the chat-based CV builder produces single-column, ATS-friendly layouts across six templates, so the structural traps are handled before you start worrying about which AI terms to include.

What are the biggest mistakes people make with AI keywords on a CV?

The costliest mistake is stuffing. A wall of thirty AI terms with no project behind them reads as noise to a recruiter and increasingly gets flagged by screening tools too, and the moment one term gets probed in a phone screen, the whole list loses credibility. Hidden white text and keyword blocks buried behind a graphic are worse: they're trivially visible in a parsed view, and they turn a formatting question into an honesty question. Here's my blunt opinion — a single bullet naming the model you used, the problem you solved and the number you moved outperforms an entire keyword cloud, every time.

The second mistake is writing about AI in the abstract. 'Leveraging AI to drive transformational outcomes' matches almost nothing in a real requirements block and tells a hiring manager nothing about what you can do on Monday. Strip the adjectives and name the mechanics: which data, which model, which tool, which measurable change. The third is applying the same CV everywhere. The same job title uses different vocabulary at different companies — one posting says 'generative AI', the next says 'LLM applications', a third says 'AI-assisted automation' — and a master CV you tailor per ad beats one polished document you blast out.

The last trap is a new one: a CV so heavily AI-written that it reads as machine-made. Identical bullet rhythms, the same three verbs, rounded percentages that feel invented — experienced recruiters now spot it in seconds, and the irony of an AI-skills CV that nobody believes a human wrote is not lost on them. Use an assistant to draft and sharpen, then put your own specifics back in: the awkward constraint, the tool you abandoned, the number you can defend in an interview. That's what makes AI keywords on a CV read as experience instead of vocabulary.

Frequently asked questions