Event Task / Objective:
To challenge participants to explore the art and discipline of prompt engineering under strict constraints.
To simulate real-world AI navigation, including restricted language, policy boundaries, and incomplete information.
To test the ability to reason, reframe, and control AI behavior through intellectually demanding tasks like debate and reverse-engineering.
To emphasize logical consistency and linguistic control over coding speed or technical jargon.
Design I Technical Specifications:
Environment: Participants will interact with AI systems to construct arguments and reconstruct prompts.
Challenge Scope:
Round 1 (Taboo Debate): Engaging in a debate with an AI opponent using a fixed stance while avoiding specific "taboo words."
Round 2 (Prompt Reconstruction): Analyzing a reference AI output and reverse-engineering the specific prompt required to generate that exact result.
General Guidelines:
Participation:
This is a solo event. All participants must be registered before the event begins on April 8th.
Resources:
Participants must use the provided AI interfaces and strictly follow the assigned debate topics and constraints.
Execution:
In Round 1, participants are limited to a specific number of prompts to encourage strategic thinking.
Submissions:
Evaluation is based on logical clarity, adherence to constraints (avoiding forbidden terms), and the accuracy/alignment of reconstructed outputs in Round 2.
Timeline:
The event runs from 2:00 PM to 4:00 PM. Participants must strictly adhere to this timeline.
Safety Rules:
Conduct: Any form of misconduct, unethical behavior, or violation of cybersecurity/AI ethics will lead to immediate disqualification.
Plagiarism: Code reuse or sharing solutions between participants is strictly prohibited and results in a permanent ban.
Integrity: Participants must operate only within the provided challenge environment and avoid actions that could disrupt the service.
Scoring & Judging Criteria
Parameter
Weightage
Max Points
Logical Clarity & Argumentation
25%
25
Adherence to Constraints (Taboo words)
25%
25
Prompt Accuracy & Alignment
25%
25
Effectiveness in AI Control
25%
25
Total
100%
100
Penalties & Disqualifications
Violation Penalty
Late Submission -10 Points
Unsafe Operation Immediate
Disqualification Incomplete or Pre-built Work -20 Points
Violation of Safety Zone -15 Seconds / -10 Points
Misconduct or Unethical Practice Permanent Ban
General Instructions:
Follow all university and MakerSpace cluster policies.
Respect judges, peers, and staff.
Any form of plagiarism or code reuse will lead to disqualification.
Decisions of the judges are final and binding.
Certificates will be awarded to all valid participants.