This document is the workspace for exploring the working groups position on the topic of “AI Generations / Creations”
Copyright and related rights are unwarranted for AI-generated outputs as AI is currently understood, for two fundamental reasons: lack of a human author and lack of originality.
First, the notion of human authorship is a bedrock principle of copyright. Direct human involvement should remain a precondition to determining whether a work is worthy of protection and whether copyright can be claimed. While the conceptualisation of AI is still in flux, the technical nature of human inputs, combined with the mechanistic nature of AI algorithms and the absence of any personality rights recognised to AI, currently provide very little ground to justify any copyright protection for AI outputs.
Second, in most cases, AI algorithms use automated, mathematical means to encode statistical information about a set of input, such as copyright works. AI uses this statistical information, combined with some random seed, to generate output which is statistically similar to or indistinguishable from an arbitrary member of the set of input works. AI-generated output is composed of snippets chosen arbitrarily from thousands or millions of input works and generated as a result of a mathematical, stochastic function. Thus, AI output should similarly be presumed to lack originality. In short, AI-generated outputs should be in the public domain, at least pending clearer understanding of this evolving technology.
The following points elaborate the position in more detail:
Copyright is not the right framework to regulate AI outputs
Copyright (including related rights and sui generis rights) is not a universal tool meant to address any and all problems that society might encounter.
Using copyright to govern AI outputs is contradictory to copyright’s primordial function of offering an enabling environment for human creativity to flourish. It should not be assumed that a system developed to protect human creative expressions can simply be applied to non-human output.
Copyright has yet to adapt to the digital environment, which emerged several decades ago; it is unwise to attempt to force the application of an already outdated system to the nascent and uncharted field of AI technology.
To the extent that copyright might be deemed relevant to regulate AI output, this policy brief provides guidance on the basic questions to consider.
No copyright for AI outputs
Although AI might arguably generate output akin to a copyright work, this output does not necessarily meet the copyright protection requirements, such as authorship and originality.
AI-generated outputs are not “works” as this term is defined in copyright law.
Granting copyright to AI output would raise important concerns around intellectual property rights overlap leading to overprotection. Other exclusive rights (including copyright and patent rights in AI software) already provide sufficient protection in the AI space. Overprotection can have negative impacts on creativity, innovation and the provision of public goods.
Introducing new rights complicates an already complex field of copyright, related rights and sui generis rights, and risks entangling the exercise of existing and any future exceptions and limitations.
AI-generated content should not benefit from any copyright (economic or moral rights), related rights or sui generis protection.
Generating content through AI is not an act of authorship
Persons who develop AI algorithms are not the authors of AI-generated outputs.
The persons (physical or legal) who program the AI should not be entitled to copyright in the output generated by the program they created.
AI-generated output is made of a mechanical conglomeration of the inputs. Absent direct human intervention or direction, this process lacks the authorial contribution necessary to warrant copyright protection.
Content generated without human creative input should be in the public domain.
AI is not capable of producing content “autonomously”
There is uncertainty about whether and to what extent AI is capable of producing content “autonomously” without any direct, material human involvement.
It is not established that AI can acquire and develop autonomous and cognitive features through experience learning that are sufficient to generate creative output. A certain level of human input is always required.
The mechanistic nature of AI algorithms bars any possibility of autonomous creation. The output of any current AI algorithm is essentially a mathematical function of its inputs, be they image or video files or textual data. This is not a valid basis for claims of autonomous authorship.
AI systems are not authors
The notion of the human creator, the human creative spirit and human creativity are bedrock principles of the copyright system.
Direct human, authorial involvement is a precondition to determining whether a work is worthy of copyright protection or whether copyright can be claimed.
Bestowing copyright upon AI-generated outputs to the copyright or patent holder of the AI software, algorithm or program, or to any of its users or to contributors of data, is not advisable absent evidence of direct human creativity and in light of concerns about over-protection.
One would be mistaken to presume an equivalence between human-authored works and AI-generated outputs. AI and the copyright principle of authorship are antithetical concepts; stated otherwise, “the very idea of AI authorship is oxymoronic.”
The generation of output through an AI process without human involvement is not an act of authorship.
AI-generated output is not likely to be considered “original”
It is not clear how to judge the originality of a work essentially composed of random snippets of thousands or millions of input works.
AI-generated outputs should not as a default be considered original works. Originality is a reflection of the intellectual, creative choices made by a human author.
Output generated solely as a result of a mathematical function lacks originality.
AI cannot have human rights
The Universal Declaration on Human Rights (UDHR) protects the moral and material interest of authors resulting from scientific, literary or artistic production and as such recognizes authors’ rights - rights deriving from human authorships as human rights.
AI systems cannot have human rights.
AI cannot be held liable for its actions
Authors are granted rights in their creations; in return, they are liable for their actions.
AI itself cannot be held liable for copyright infringement nor for any other breaches of rights caused by its generated output, due to a current lack of regulation of AI liability. Since “copyright (as in a right in one’s work) and responsibility for that work historically have gone hand in hand,” AI-generated output should not be afforded copyright protection.
AI development does not require copyright in AI-generated output as an incentive
Producers of AI do not require copyright protection in AI-generated outputs as an incentive to develop AI.
Market opportunities are already rife, in the absence of any copyright protection for AI-generated outputs.
The efforts invested in the creation or invention of the AI program are to be incentivized are rewarded through means other than copyright protection over the AI-generated output. Incentives and rewards in recognition of the investment made and the innovation brought about by the organizations and individuals involved in the development of AI can be found in other areas, including copyright in the code of the AI software itself, as well as patents, trade secret laws, and laws protecting against unfair competition.
Copyright’s utilitarian doctrine and incentives theory cannot support a claim that AI be afforded rights for any generated output because AI fails to meet the role of the author and its contribution to human-led social progress. Granting AI-outputs the status of copyright work goes against the social purpose for which copyright was created
Very limited protection for AI outputs as a fall-back option
Assuming the undesirable outcome that new copyright, related rights or sui generis rights would be established to regulate AI and protect AI-generated content despite our strong opposition, this should be done conservatively and with restraint.
Considerations should include: a high bar for the creation of such new rights, a short term of protection, and robust exceptions and limitations to uphold users’ rights, safeguard the public interest, and ensure a vibrant public domain.
No new sui generis rights for AI-generated content
There should be no new sui generis rights established for AI-generated content.
The incentives that already exist within the copyright (for AI software) patent, trade secrets, and unfair competition systems are sufficient to encourage the development of AI applications.