MOTION: AI Content Policy PASSED 5-0-1

[Jubal]

This is a binding vote of the Plaza on the below policy. Simple majority, one week to vote.



1. The plaza recognises:

• That images, text, video, and other media produced by large generative adversarial models, more commonly known as "AI" and hereafter referred to as generative AI, are increasingly present in the fields of writing, academia, game design, and other creative and intellectual efforts.
• That media produced by such models are potentially injurious to small and hobbyist creators, whose work may be scraped and partially replicated by such models without their permission.
• That the potentially very rapid generation of such media may risk drowning out or obfuscating discussion of small-scale creative activities by flooding information infrastructure with automatically generated content.
• That the legal status of generated works is still unclear, especially where large models have been trained on significant bodies of copyrighted works.
• That other impacts of using generative AI content may have wider societal harms, including but not limited to the high environmental costs of server infrastructure, the misinformation potential of generative AI information summaries, and the replication of existing biases through generative AI.
• That owing to the nature of this content and the potential for partial use, it may not always be possible to determine whether generative AI has been used in a given project or situation, and that mechanisms for making such a determination may themselves be flawed or biased.
• That generative AI is being marketed as a potential solution to many challenges in creative projects, and in certain cases creators may adopt and use it without fully understanding its operation or associated problems.

2. The plaza believes:

• That as an organisation we exist to promote and support creators, and should take measures to avoid the promotion of systems that may be injurious to artists, writers and musicians in our community.
• That retaining a cautious approach to potential legal risks around copyright is sensible to ensure the continued functioning of Exilian.
• That it is nonetheless important to retain capacity for open discussion and exploration of new technologies.

3. The plaza resolves:

• That retaining free expression and speech are important, and therefore discussion of, and limited posting of, AI generated content is broadly acceptable on Exilian, especially for discursive and exploratory reasons.
• That the use of AI generated content should be disclosed, and that if an intentional failure to make such a disclosure occurs then the material may be removed and the action of posting it treated as a form of spam under Exilian's Terms of Service.
• That excessive posting of generative AI content, for example primarily posting model-generated posts or posting projects where a large percentage of the art or writing is explicitly generative AI driven, may be treated as a form of spam under Exilian's Terms of Service.
• That any generative AI content posted must not be used to mimic the work of specific writers or artists, or the look of real individuals: such content should be removed to avoid legal challenge, and may be considered under the piracy clause of the Terms of Service.
• That Exilian's newsletters and social media should avoid the promotion of generative AI and any projects that are known to incorporate it.
• That Exilian's competitions and events should have in future explicit rules against the submission of generative AI works.
• That the above clauses 3.3, 3.4 and 3.5 may be waived if the model used is both open source and can be shown to have been entirely trained on public domain/CC-0 material or material where the model's producer is the creator and copyright holder.
• That clause 3.5 can be waived for projects which have used AI generated content in the past but have since removed or replaced it. These may be reviewed on a case-by-case basis.
• That Exilian will comply with legal obligations regarding generative AI content, and that such content may be removed at any time if the administrators have reason to believe that it poses any legal risk to the functioning of Exilian.
• That Exilian will attempt to take any reasonable measures to prevent data scraping of users' content for AI purposes, and that such data scraping will be treated as in violation of Exilian's terms of service, but that Exilian remains a public website run by volunteers and as such cannot provide guarantees that such scraping will be blocked in all cases or legal support in cases where scraping may have taken place.
• Exilian shall not publicly encourage or signpost users to generative AI which does not meet the stipulations in 3.7 as a solution to ongoing challenges in their projects, and shall endeavour to improve support and accessibility for alternative sources of creative material.

[Pentagathus]

DOWN WITH THE MACHINES!

[Jubal]

Motion PASSED 5-0 with 1 abstention.

[Jubal]

This motion has been amended by a separate motion of the Plaza, carried six votes to zero with two abstentions, which contained the following additions:

a) An addition to section 1, "The plaza recognises" as follows:
1.7 That generative AI is being marketed as a potential solution to many challenges in creative projects, and in certain cases creators may adopt and use it without fully understanding its operation or associated problems.

b) A modification to 3.5:
The substitution of the word "clearly" with "are known to".

c) An extra clause:
3.8 That 3.5 can be waived for projects which have used AI generated content in the past but have since removed or replaced it. These may be reviewed on a case-by-case basis.

d) Another extra clause:
3.11 Exilian shall not publicly encourage or signpost users to generative AI which does not meet the stipulations in 3.7 as a solution to ongoing challenges in their projects, and shall endeavour to improve support and accessibility for alternative sources of creative material.

A copy of the rules before the amending motion was carried can be found in spoiler tags below:

Spoiler

1. The plaza recognises:

• That images, text, video, and other media produced by large generative adversarial models, more commonly known as "AI" and hereafter referred to as generative AI, are increasingly present in the fields of writing, academia, game design, and other creative and intellectual efforts.
• That media produced by such models are potentially injurious to small and hobbyist creators, whose work may be scraped and partially replicated by such models without their permission.
• That the potentially very rapid generation of such media may risk drowning out or obfuscating discussion of small-scale creative activities by flooding information infrastructure with automatically generated content.
• That the legal status of generated works is still unclear, especially where large models have been trained on significant bodies of copyrighted works.
• That other impacts of using generative AI content may have wider societal harms, including but not limited to the high environmental costs of server infrastructure, the misinformation potential of generative AI information summaries, and the replication of existing biases through generative AI.
• That owing to the nature of this content and the potential for partial use, it may not always be possible to determine whether generative AI has been used in a given project or situation, and that mechanisms for making such a determination may themselves be flawed or biased.

2. The plaza believes:

• That as an organisation we exist to promote and support creators, and should take measures to avoid the promotion of systems that may be injurious to artists, writers and musicians in our community.
• That retaining a cautious approach to potential legal risks around copyright is sensible to ensure the continued functioning of Exilian.
• That it is nonetheless important to retain capacity for open discussion and exploration of new technologies.

3. The plaza resolves:

• That retaining free expression and speech are important, and therefore discussion of, and limited posting of, AI generated content is broadly acceptable on Exilian, especially for discursive and exploratory reasons.
• That the use of AI generated content should be disclosed, and that if an intentional failure to make such a disclosure occurs then the material may be removed and the action of posting it treated as a form of spam under Exilian's Terms of Service.
• That excessive posting of generative AI content, for example primarily posting model-generated posts or posting projects where a large percentage of the art or writing is explicitly generative AI driven, may be treated as a form of spam under Exilian's Terms of Service.
• That any generative AI content posted must not be used to mimic the work of specific writers or artists, or the look of real individuals: such content should be removed to avoid legal challenge, and may be considered under the piracy clause of the Terms of Service.
• That Exilian's newsletters and social media should avoid the promotion of generative AI and any projects that clearly incorporate it.
• That Exilian's competitions and events should have in future explicit rules against the submission of generative AI works.
• That the above clauses 3.3, 3.4 and 3.5 may be waived if the model used is both open source and can be shown to have been entirely trained on public domain/CC-0 material or material where the model's producer is the creator and copyright holder.
• That Exilian will comply with legal obligations regarding generative AI content, and that such content may be removed at any time if the administrators have reason to believe that it poses any legal risk to the functioning of Exilian.
• That Exilian will attempt to take any reasonable measures to prevent data scraping of users' content for AI purposes, and that such data scraping will be treated as in violation of Exilian's terms of service, but that Exilian remains a public website run by volunteers and as such cannot provide guarantees that such scraping will be blocked in all cases or legal support in cases where scraping may have taken place.