Summary

  • APRICOT's 2027 Fellowship rules require applications in the applicant's own words and prohibit AI tools, chatbots or LLMs from writing or substantially drafting them.
  • The selection criteria ask for specific, verifiable accounts of hands-on Internet operations and active local technical-community work, not generic job descriptions or passive membership.
  • Applications close on 12 October at 23:59 Hong Kong time. The public material does not describe AI detection, a verification process or any applicant outcome.

The application page draws an unusually direct line between a Fellowship claim and the person making it. Applicants must write their own applications; AI tools, chatbots and large language models—including ChatGPT, Claude and Gemini—must not write or substantially draft the text. Generic or AI-generated prose, the page says, will count against an application because it may not reflect the applicant's actual experience.

That instruction is part of the APRICOT 2027 Fellowship call publicized by APNIC Blog on 28 September. The official key-date page puts the closing time at 23:59 Hong Kong time on 12 October. Reviews are scheduled for 13–25 October, a clarification period for 26 October–5 November, and invitations from 6 November. The Fellowship Committee's terms say its decisions are final and non-negotiable.

The rule is not standing alone. Eligible applicants must live in a developing Asia-Pacific economy under the Australian Government's definition and be active in Internet technical operations. The published examples include network engineering, network operations and systems engineering. Applicants must describe their local Internet-community work, demonstrate relevant expertise for any selected advanced training, and explain financial hardship that would otherwise make attending impossible.

The Committee says it particularly values detailed descriptions of day-to-day hands-on work over broad job titles; concrete involvement in a local NOG, IXP, CERT or similar group rather than passive membership; and a personal, specific explanation of why self-funding or employer sponsorship is not possible. These are requirements for an account of experience, not a published list of documents that applicants must attach. The sources reviewed do not show a scoring formula or an evidentiary checklist.

There are three separate questions here. The Fellowship Management System says applicants access its portal with APNIC Login. That is the account gate. The own-words condition concerns who composed the application. The Committee then decides whether the applicant meets its published criteria. A login does not, on the available record, establish authorship, operational experience or selection—and the public pages do not say that the portal verifies any of those things.

That distinction matters in both directions. The explicit rule should not be diluted into a vague request to “use AI responsibly”: the page says not to use AI tools to write or substantially draft an application. But it would also be wrong to claim that the Committee runs AI-detection software, rejects a particular kind of applicant, or has already found a violation. No application, decision, enforcement method or appeal process was published in the sources reviewed. Nor does the page resolve every borderline case, such as which non-drafting assistance falls within its wording.

It directs application questions to the Fellowship support contact; it does not invite applicants to test the boundary by submitting generated prose.

The practical test described is narrower and more human: can an applicant explain a real operational responsibility, show how they participated in a local technical community, and give details that can be checked? A polished generic account is not a substitute for that evidence. The rule places authorship with the applicant; the selection decision remains with the APRICOT Fellowship Committee.

Sources