This page is the canonical version of XTF's competition guidelines for the 2026 cycle. If anything below conflicts with text on another page, this page wins.
1 · Submission requirements
To be eligible for review, your team must upload all of the following to its team dashboard by 31 August 2026, 23:59 UTC:
Project paper (PDF) — up to 12 pages including figures, references and appendices. Use the XTF paper template (LaTeX or Word) or any reasonable approximation. A paper is required for every project, including pure hardware or software projects.
Short video (MP4 or YouTube/Vimeo link) — up to 5 minutes. The video should explain what you did, why it matters, and demonstrate any working prototype. Production value is not scored; clarity is.
Repository or supporting materials link — a public link to your code, design files, datasets, or any other artefact a reviewer would need to evaluate the work. Private repos with read-only review access are fine.
Team statement (≤500 words) — who did what, and (if applicable) what mentorship or external help you received. Honesty here is required and is itself a positive signal.
Ethics & safety attestation — a signed checklist (provided in your dashboard) covering human subjects, biosafety, and youth safeguarding where applicable. See section 6.
Languages
Submissions are accepted in English, Spanish, French, Mandarin, Arabic and Portuguese. If you write in another language, you may submit but must also include an English-language abstract (≤300 words) and an English-language version of the 5-minute video (subtitles are fine). Reviewers are matched to your submission language whenever possible.
2 · File & format specs
Item
Format
Max size
Paper
PDF/A preferred
25 MB
Video
MP4 / WebM / public URL
500 MB (file); ≤5 min
Code / data
Public Git repo, Zenodo, OSF, or similar
Link, no limit
Team statement
PDF or plain text
2 MB
Ethics attestation
Signed PDF (dashboard generated)
2 MB
3 · Judging rubric
Every submission is scored on five axes. Each axis is rated 1–5; the team's score is the weighted average. The rubric is published in full below; reviewers must justify each score in writing.
Criterion
Weight
What "5" looks like
Scientific rigor
25%
Well-posed question, appropriate method, honest treatment of uncertainty, reproducible analysis.
Originality
20%
Adds a result, method or framing the field did not already have; clearly situates itself in prior work.
Impact & relevance
20%
If this worked at full scale, it would change something important. Problem is worth solving.
Execution & craft
20%
Prototype works. Code runs. Data is well-organized. Documentation lets a third party reproduce.
Communication
15%
Curious non-specialist can follow the work without sacrificing rigour. Figures earn their space.
4 · Review process
Intake check (1 week). XTF staff verifies that submissions are complete and within policy.
Blind first round (3 weeks). Names, schools, and countries are stripped. Each project is scored by ≥3 independent reviewers. Two of three must agree the project advances.
Shortlist (1 October). Top 5% per track is announced publicly. Every entrant — shortlist or not — receives written reviewer feedback.
Semi-finals (7 November). Shortlisted teams present live to a track panel in 20-minute slots (10-minute talk, 10-minute Q&A). Reviewer panels are 4–6 working scientists.
Grand Finals (5–6 December, Geneva). 24 finalists (3 per track) present in person. Final scoring combines paper, prototype demo and in-person Q&A.
Awards (6 December). Grand Champion, Track Winners and Division Winners announced.
5 · Intellectual property & data policy
Short version
You own your work. XTF claims no IP in your project. To be published in the XTF Proceedings (optional, opt-in), your paper must be licensed CC BY 4.0. Code and data may be licensed however the team chooses, but open licenses are weighted positively in the rubric.
Long version:
XTF acquires no ownership of your submission. We acquire only a non-exclusive, non-commercial license to display your paper, video and a still image for purposes of running and promoting the competition.
If your project advances to the Shortlist, your abstract may appear on xtf.org. You can opt out.
If your project is invited into the XTF Proceedings, you must license that paper under CC BY 4.0. You may decline the invitation.
Personal data submitted by participants (name, age range, country, contact) is handled per XTF's Privacy Notice and is never shared with third parties for commercial use.
6 · Ethics & safety
Human subjects. Any research involving people requires informed consent. For under-18 subjects, parental/guardian consent is required. School- or university-level ethics review is required where applicable.
Biosafety. Wet-lab biology must follow the biosafety guidance of the host institution and be supervised by a qualified adult. Pathogenic organisms, gain-of-function research, and any work outside BSL-1/BSL-2 norms are not permitted.
Animal subjects. Vertebrate animal work must follow the institution's animal-care protocols and is restricted to non-invasive observation.
Dual-use. XTF will not accept any project whose primary purpose is to cause harm — including offensive cybersecurity demonstrated against real non-consenting systems, weapons design, or surveillance systems sold to states without judicial oversight.
Youth safeguarding. All mentors, judges and staff who interact with under-18 participants pass a background check appropriate to their jurisdiction. Direct messaging between adults and under-18 participants must go through XTF's monitored channels.
7 · Use of AI tools
AI tools are part of modern science. We don't ban them — we ask you to be honest about them.
You may use AI assistants for ideation, writing, coding and analysis. Disclose what you used and how, in your Team Statement.
You may not submit work that was substantially generated by an AI system and not understood by your team. If a judge asks "explain this code" or "explain this paragraph," you must be able to.
Models you trained or fine-tuned for the project are obviously fine and welcome. Document them.
For the AI & ML track specifically, undisclosed reliance on third-party models is a disqualifying offence.
8 · Code of Conduct
XTF events — virtual or in person — are intended to be respectful, curious and safe for everyone. The full Code of Conduct is posted at xtf.org/code-of-conduct. The short version:
Treat every other participant, mentor, judge and staff member with respect, regardless of their age, gender, nationality, language, religion, ability or background.
Harassment in any form — including unwelcome contact, slurs, stalking, doxxing or threats — is grounds for immediate removal.
Plagiarism, data fabrication or claiming someone else's work as your own results in disqualification and is reported to the team's home institution.
Substance use (including alcohol) is not permitted at any XTF event for under-18 participants. Adult-only XTF events follow the host country's laws.
If you experience or witness a violation, contact safe@xtf.org. Reports are received by XTF's Safeguarding Officer, not by your track judges.
9 · Grounds for disqualification
Plagiarism or data fabrication.
Submission of work outside the eligible cycle window (before 1 March or after 31 August).
Team-size violations (more than four members; one participant on multiple teams; one team submitting multiple projects).
Failure to disclose substantial AI assistance, or failure to explain submitted work at the judging round.
Material false statements on the registration or ethics attestation.
Violations of the Code of Conduct judged by the Safeguarding Officer to warrant removal.
10 · Appeals
If you believe a decision was made in error or process, you may appeal in writing to the Director of Judging & Review (appeals@xtf.org) within 10 days of the decision. Appeals are reviewed by a three-person panel that includes at least one board member not previously involved. The panel's decision is final.
This document is version 1.0 of the 2026 Guidelines. The Board of Directors may issue clarifying addenda; any change after registration opens will be limited to clarification, not substantive policy change.
Subscribe
Be the first to know.
Get track-brief releases, deadline reminders and clarifying addenda the moment they're published.