Submission 112
Spinifex Press Pty Ltd
Submitter group: Creative sector.
An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.
Summary
Read the original submission for the full text.
An independent publisher and author argues that AI offers no opportunities for writers, publishers or small business and that AI training on books without permission is theft. Describes pirated titles, image-based abuse and data-centre impacts, and recommends stronger copyright rules, AI disclosure, no copyright in AI-generated work and payment deductions for AI content.
Key points
Quotes are the submitter’s own words. Text taken from documents, including scanned pages read by text-recognition software, can differ from the PDF. Check the original. Listed in the same order as on the submission’s page in the interactive guide.
Argues that scraping works for text and data mining and large language models without permission is theft.
To scrape my work and use it in Text and Data Mining (TDM) without my permission is theft. This is precisely what the big tech companies are doing with TDM and Large Language Models (LLMs).
Says it found 145 of its books, almost half its list, among the copied works on the list published by The Atlantic.
When The Atlantic list was published, we found 145 books that had been copied. This is almost half of our list. This represents 34 years’ work by 5 employees, 2 freelancers and 2 company owners.
Argues that home-grown AI offers no benefit to Australian publishers and writers, since theft has become the norm in the AI sector.
There is no benefit for Australian publishers and writers to have home-grown AI. Since theft has become the norm in the AI sector, it is highly unlikely that Australian AI companies will behave any differently.
Describes the increasing sexualisation of women through AI as a policy concern.
The increasing pornification of women through AI seems like it might be someone else’s problem.
Describes concerns about data centres' ecological impacts given Australia's electricity and water constraints.
One of my concerns about AI is the ecological impact of data centres on the environment in a country which does not have national infrastructure for electricity and has a severe shortage of water resources.
Proposes that author payments through Copyright Agency, Public Lending Right and Educational Lending Right be reduced in proportion to AI-generated content.
Payments due to the author – as in the Australian organisations Copyright Agency, Public Lending Rights, Educational Lending Rights – should be dealt with as above: 5% deducted if 5% is AI generated; 50% deducted if 50% is AI generated; no payment to works that are 100% AI generated.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Enforce Community Intellectual Rights and Traditional Resource Rights protecting collective creativity.
Some creativity is undertaken by groups: group devised performances; first nations arts practice; national cultural property whether 65,000 years old or created now. These are potentially protected by Community Intellectual Rights (CIRs) and Traditional Resource Rights (TRRs), but these rights need to be enforced.
Strengthen copying laws to prevent fair-use claims, while allowing short extracts from books over 50,000 words and poetry or song extracts up to five lines and no more than 5% of the poem.
Strengthen laws around copying so that no one can claim fair use. Allow for the copying of short extracts from books (of over 50,000 words), some copying of a few lines of poetry or songs (up to 5 lines, but not more than 5% of the total poem).
Make AI copying illegal where an author or publisher does not allow use for training.
Make AI copying illegal if an author or publisher does not allow use for training.
Require all work that uses AI to disclose the extent of AI involvement.
To protect transparency and trust, all work that uses AI should disclose the extent of AI involvement.
Make it impossible to copyright AI-generated work.
Make it impossible to copyright AI generated work.
Clearly mark any AI-generated sections of a work.
Any AI sections of the work should be clearly marked as AI generated.
Deduct author payments (Copyright Agency, Public Lending Rights, Educational Lending Rights) in proportion to the AI-generated share of a work, with no payment for works that are 100% AI generated.
Payments due to the author – as in the Australian organisations Copyright Agency, Public Lending Rights, Educational Lending Rights – should be dealt with as above: 5% deducted if 5% is AI generated; 50% deducted if 50% is AI generated; no payment to works that are 100% AI generated.
Subject individuals or institutions flouting the AI-generated payment-deduction rule to the same rules as for copyright breaches.
5% deducted if 5% is AI generated; 50% deducted if 50% is AI generated; no payment to works that are 100% AI generated. Individuals or institutions flouting the rule be subjected to the same rules as for copyright breaches.
Run campaigns, like those run against smoking, if governments are serious about protecting the intellectual property rights of creative people.
If governments were serious about protecting the intellectual property rights of creative people, they would be running campaigns like those run in Australia against the smoking of cigarettes.
Consider Traditional Resource Rights, or another system benefiting those most likely to be dispossessed by Western IPRs, to protect Indigenous and traditional knowledge.
TRRs, or some other system which is developed to benefit those most likely to be dispossessed by western IPRs, could create the conditions for a fair way of proceeding with the protection of Indigenous and traditional knowledge.