STEP is built so faculty, university leaders, and learners can understand what is being assessed. The rubric is published, the simulation logic is explainable, and the evidence model is designed for institutional use.
The four STEP dimensions give a practical language for AI-era work: can the learner frame the problem, test the AI output, explain judgment, and deliver something useful?
Turns an ambiguous brief into a clear plan, including what is missing, what must be clarified, and what assumptions are safe to make.
Checks AI-generated work for weak evidence, hallucinated claims, missing context, and logic gaps before passing it forward.
Explains thinking, is honest about AI use, handles pushback, and keeps people aligned.
Delivers usable work under time pressure, with incomplete information and realistic trade-offs.
STEP uses three performance levels: Developing, Competent, and Proficient. A credential requires Competent or above across all four dimensions.
Singapore's higher-education direction points to a common AI foundation for students, discipline-specific application, and human judgment at the centre. STEP is built for the next question: can students turn AI competency into work a manager would trust?
Singapore's public direction points toward baseline AI competencies, field-specific use, and responsible guardrails across higher education.
The NUS coverage stresses confidence with AI alongside curiosity, critical thinking, adaptability, and judgment.
Global workforce research points to AI and data growth while analytical thinking, resilience, leadership, and collaboration remain critical.
STEP combines learning science, real-work assessment, and workplace judgment design. The purpose is not to reward smooth AI output; it is to show whether the learner can use AI while preserving evidence, accountability, and trust.
Students face a realistic situation, see consequences, receive feedback, and improve on fresh variants.
The evidence comes from briefs, audits, disclosures, revisions, recommendations, and handoffs.
Students must respond to incomplete instructions, people pressure, realistic AI errors, and hard choices.
The credential rests on repeated observable actions across the mission path, not a single answer.
Each task starts with the capability claim, the evidence needed, and the scenario moment that can reveal it.
Students must show what AI did, what they verified, what they changed, and what must remain human-owned.
Each capstone places the learner inside a realistic workplace situation where AI is useful but imperfect. The structure is consistent enough to compare cohorts and realistic enough to reveal judgment.
STEP supports readiness, advising, curriculum, and employability conversations by producing a clear simulation record. Group evidence and faculty reviews are used to refine scenario design, example answers, and intervention priorities over time.
The current claim is intentionally specific: a learner demonstrated competence in a STEP simulation using a published standard. Hiring prediction, market norm, and accreditation-compliance claims require separate proof.
STEP certification is issued by AIR APAC using the STEP rubric. Exeter Labs develops the simulation and reporting platform.
If you want to understand the standard quickly, run the public demo and then read the scorecard.
These sources inform STEP's policy and methodology framing. They do not imply endorsement, compliance, or accreditation.
STEP is aligned with the direction of Singapore's 2027 higher-education AI competency move. It is not a Singapore-government accreditation, compliance certification, or claim of endorsement by any named university.