The rise of generative AI forces a legal and ethical reckoning over copyright, originality, and the definition of human authorship in creative works.
GENERATIVE artificial intelligence (Gen AI) is more than just a tool; it is a “collaborator” that is slowly redefining originality, creativity and innovation.
We use Gen AI in just about anything these days; from automated design, content generation to predictive analytics, machines have begun producing outputs that mirror human work.
However, deeper questions linger: To what extent can a human creator claim originality and ownership on Gen-AI outputs?
This is a multi-faceted dilemma. It is not just a technical and legal issue but also a highly ethical one as IP laws and frameworks were established on the foundation of human ingenuity. Legal systems are now faced with the urgent need to adapt to the fast-evolving industry enabled by AI and machine learning.
While specific regulations and guidelines for allowing copyright registration of Gen-AI assisted works are yet to be established, the World Intellectual Property Organisation and World Trade Organisation, through their Trade Related Intellectual Property Rights Agreement, maintain that only human creators are recognised as authors currently. This consistency is observed across the jurisdictions of member states including Malaysia.
Risks of Gen AI: Are they worth it?
In the US class action lawsuit Bartz v Anthropic (2024-2025), a trio of non-fiction authors Andrea Bartz, Charles Gaeber and Kirk Johnson sued Anthropic for training its AI model Claude to use their books without permission.
Anthropic pirated copyrighted books from Library Genesis and Pirate Library Mirror for training Claude initially, but “remodelled” Claude’s training using cleaned-up, digital scans of legally acquired books. A partial summary judgment was given, stating that training Claude using the legally acquired books was transformative fair use, but stressed that Anthropic has no entitlement to use the millions of pirated copies in the AI model’s central library.
Although Anthropic agreed to destroy all pirated copies of books they previously acquired, the ruling for the present case does not free Anthropic of future alleged input and output infringement as a result of training Claude’s model. In fact, Anthropic is currently facing multiple lawsuits from several major record labels.
The lesson for us is that we should give the claim of originality and authorship of Gen AI-assisted works under copyright law, and by extension, other IP laws a serious treatment as the resulting reputational and financial damages are hefty.
Using Gen AI responsibly to prime creativity
The message is clear: If Gen-AI outputs are developed without meaningful and substantial human inputs, the outputs fall outside the scope and rights provisioned by the copyright law.
While there is a need for the law to rapidly adapt to Gen AI, we must first differentiate between AI-assisted and AI-generated outputs. For AI-assisted outputs, AI is used as a tool to aid human creators in the creative process. This contrasts with AI-generated outputs, where AI demonstrates creative abilities independent of the human creator.
Legal regulations and guidelines must be established around AI-assisted outputs. The degree of intervention to transform AI-generated outputs must be determined to assess the degree of reliance on the AI tool in the entire creative process. In other words, the burden of proof lies with the human creator to demonstrate that Gen-AI outputs are a result of series of tasks of at least one creative process with substantial human intervention.
When producing AI-assisted outputs, human creators should maintain clear logs and records of their contributions, such as the intent behind prompts, the selection and editing of generated content and the creative decisions that guided the process. These records not only support claims of originality but serve as evidence in the event of legal scrutiny or disputes.
In this evolving landscape, the path forward is about conscience and compliance in equal measures. Creators and organisations must take a proactive and principled approach to Gen AI: be informed, understand what the law allows, and use Gen AI ethically and transparently. It is the human touch – our experience, skills, intent, values and judgement – that ultimately defines creative expression.
Nor Farah Natasha Tajuddin is an ecosystem builder based at Taylor’s University’s Research and Enterprise Department: Knowledge Transfer and Commercialisation. She is an IP specialist certified by the Intellectual Property Corporation of Malaysia and a registered technology commercialisation associate of the Malaysian Research Accelerator for Technology and Innovation. Comments: [email protected].
