The UN Women evaluation internship offers six months of full-time experience in New York. The role combines evaluation work with responsible AI use. Applications close on 13 November 2026, with conditional stipend support available.
Job identification: 37837. Source checked: 9 November 2026.
UN Women evaluation internship: what the opportunity involves
The work supports AI-assisted evaluation, evidence checking, data analysis and communication. Related university studies and data-analytics experience are required; practical generative-AI experience is preferred.
Eligibility requirements
Eligible routes include graduate students, final-year bachelor’s students, graduates able to start within two years of graduation, and qualifying postgraduate traineeships linked to a degree. English is required. Review all functional competencies in the official listing.
Stipend and practical conditions
Interns without financial support from another institution receive a stipend that partly supports living costs; the amount is not stated. Successful applicants must provide required education, identity, visa where applicable and valid health-insurance evidence. No remote arrangement or full-funding package is confirmed.
How to apply
Complete every required application section in the recruitment portal. The displayed closing time is 20:53 on 13 November, but its timezone was not established. Check the portal’s display and submit early rather than assuming a local cutoff.
Prepare evidence of responsible AI work
The following suggestions are independent editorial advice, not extra employer requirements or a description of a confirmed recruitment exercise.
Choose an example you can explain from beginning to end
Select a real analytical or research task where you used a tool and checked the result. Explain the original question, the information available, your contribution and the limitations. A small project you understand thoroughly is more useful than a large claim supported only by screenshots.
Be clear about whether the work was individual, collaborative or supervised. Distinguish a method you practised from one you only read about. If you used an existing tool or template, explain what you adapted rather than claiming you built the whole system.
Show how you verified an output
Prepare to discuss a time you checked an automated result against its source. Describe how you noticed an error, omission or ambiguous statement and what you did next. Do not use a tool’s confident wording as evidence that its answer was correct.
You can explain verification without sharing private source material. Use public or properly authorised examples, and remove confidential information from samples. Never upload another organisation’s sensitive records to a public tool simply to create application evidence.
Separate findings from interpretation
When describing an analytical project, identify what the data showed and what you concluded from it. Mention assumptions that affected the result. Avoid presenting a correlation, a model output or a small sample as proof of a broad causal claim.
If your project used generated text, state where human review was needed. A useful example explains your judgement, including a decision not to rely on an output, rather than treating automation itself as the achievement.
Demonstrate careful communication
Think of a report, presentation or briefing where you adapted technical material for a particular audience. Explain how you chose what to include, made uncertainty visible and helped the reader understand the main point.
Check any permitted sample for accessible labels, readable figures and clear source attribution. Do not use decorative design to hide a weak explanation. If a document is not requested, describe the work in your application rather than adding unnecessary attachments.
Keep academic and experience dates precise
Check your enrolment status and graduation date against the eligibility route you intend to use. If a certificate is pending, describe that accurately. Do not assume that being eligible on the application date is enough when a condition concerns the start date.
List relevant experience with consistent dates and honest descriptions. If you have completed coursework involving AI, distinguish that from professional deployment or independent research. The application should make your current level of experience understandable.
Assess the practical terms before accepting
Ask for the stipend terms, start date, attendance arrangements and required documents if you progress. Estimate New York living costs and identify expenses that would remain your responsibility. Do not substitute figures from another UN internship for a written offer.
Retain the final application and any submission confirmation. Review the live deadline display carefully and leave time for technical problems. An account profile or an incomplete form is not evidence that the vacancy application was submitted.
Frequently asked questions
Should I call myself an AI expert after a short project?
Describe the actual task, tools and level of responsibility. Specific evidence is stronger than an inflated label.
Can I share confidential data as a work sample?
Only share material you are authorised to disclose. Use an appropriate public or anonymised example when possible.
Is the displayed deadline time definitely New York time?
That timezone was not confirmed. Check the live portal instead of assuming.
The UN Women internship programme guide gives background on the wider programme. This New York evaluation vacancy has its own specific criteria and deadline.
Official source and disclaimer
Read the official UN Women vacancy. This independent guide is not a recruitment service and does not guarantee shortlisting, appointment or later employment.
One thought on “UN Women Evaluation Internship 2026: Gain Six Months of AI Experience with Conditional Stipend Support”