IP-Safe AI Art Hackathon
Context
Clear, comprehensive, and age-appropriate rules and participant agreements are foundational to running an IP-safe and ethically sound AI art hackathon for minors. These documents set expectations, define boundaries, allocate rights and responsibilities, and provide a framework for enforcement.
Essential "Play Rules": Setting Expectations
The core hackathon rules should directly address the unique challenges posed by generative AI, translating complex legal principles into practical guidelines for young participants.
Human Authorship Standard: The rules must explicitly state that submitted works require significant human authorship beyond simply generating an image from a prompt. Reference the USCO's standard: AI can be used as a tool, but the human participant must be the "mastermind" exercising creative control over the final work. Provide concrete examples:
Acceptable: Using AI to generate initial concepts then substantially modifying them through digital painting; selecting and creatively arranging multiple AI elements into a unique collage; using AI to apply a style to the participant's own original sketch or photo..
Unacceptable: Submitting an image directly from an AI generator with minimal or no changes; making only minor edits like cropping or basic color adjustments to an AI output. This rule directly confronts the core copyrightability issue and sets a clear standard for participant effort.
Prompting Guidelines: To mitigate infringement risks associated with inputs, the rules must prohibit specific types of prompts:
Prompts requesting generation "in the style of" a specific, named contemporary artist whose work is likely protected by copyright (e.g., "in the style of Yayoi Kusama"). Explain briefly that while style itself isn't copyrightable, such prompts increase the risk of the AI generating infringing elements.
Prompts referencing specific copyrighted characters (e.g., "Pikachu"), logos (e.g., "Nike swoosh"), or well-known artworks still under copyright.
Prompts involving the upload of reference images unless the participant affirms they created the image themselves or have explicit permission to use it.
Prompts containing any personally identifiable information (PII) – such as names, addresses, school names – or any other confidential or private data. Explain this is for privacy and safety. These guidelines target known infringement pathways and critical data privacy risks.
Prohibited Content: Include standard clauses forbidding the creation or submission of content that is illegal, obscene, defamatory, harassing, discriminatory, or otherwise harmful or offensive. Explicitly mention the prohibition of using AI to create harmful deepfakes, non-consensual intimate imagery, or content promoting hate speech or dangerous activities. This addresses broader safety and ethical obligations, particularly crucial when working with minors and powerful generative tools.
Tool Usage: Clearly specify the policy on AI tools. Options include:
Mandating specific tools: Requiring participants to use only pre-vetted and approved AI platforms provided or recommended by the organizers. This offers more control over potential risks associated with different tools' training data or terms.
Allowing participant choice with responsibility: Permitting participants to use any generative AI tool but explicitly stating they are responsible for reviewing and complying with that tool's terms of service, privacy policy, and any restrictions on output usage. Whichever approach is chosen, it should be