Our Publication

Generative artificial intelligence (AI) – from text and image generators to music composition and even invention algorithms – is transforming how creative and innovative works are produced. But this rapid advance poses a fundamental question: Can Indonesia’s current intellectual property (IP) laws cope with AI-generated creations?In particular, Indonesia’s copyright and patent regimes face challenges around authorship, originality, and inventorship when a “non-human” intelligence contributes to or wholly creates a work. This article explores how Law No. 28 of 2014 on Copyright and the patent law (as amended through Law No. 13 of 2016 and its 2024 amendment) address – or fail to address – generative AI outputs, and what changes may lie ahead. We’ll also compare approaches in other jurisdictions like Australia and Russia to see how they are grappling with AI in the IP arena.

Copyright in the Age of AI: Authorship and Originality Under Indonesia’s Copyright Law (Law 28 of 2014), two key criteria must be met for a work to receive protection: originality and human authorship. First, a work must be “unique and personal,” resulting from the inspiration, creativity, skill or expertise of its author. Second, the law defines an “author” (pencipta) as “a person or several persons who individually or jointly produce a unique and personal work.” In Indonesian legal doctrine, “person” encompasses natural persons (human individuals) or legal entities​. An AI system is neither – it’s essentially software created by humans​.

This creates a fundamental hurdle: a fully AI-generated poem, painting, or music track might be original in a colloquial sense, but if no human intellect shaped the expression, does it meet the Indonesian law’s originality test? And without a human creator, can there be a legally recognized author? Indonesian authorities have indicated the answer is no under current law. Works generated solely by algorithms with “little to no direct involvement of a human being (aside from entering text prompts)” likely do not qualify as protectable creations in Indonesia. Likewise, because AI cannot be a “natural or legal person,” it cannot be an author under the statute. In short, a completely AI-produced output “fails to satisfy both requirements” of Indonesian copyright protection​..

Who Owns AI-Generated Works? (Authorship, Ownership and Licensing) If a creation doesn’t legally qualify as a “ciptaan” (work) under the Copyright Law, it enters a grey area. Authorship and ownership of such AI-generated material become uncertain. By default, if no human author can be identified, the work may fall into the public domain – meaning no one can claim exclusive rights. For businesses, this is a serious concern: imagine investing resources to have AI produce marketing copy, designs or software code, only to find you hold no enforceable copyright over the output. Indeed, if AI-made works aren’t protected, companies “would not be able to claim ownership over the works” (texts, images, music, etc.) created with Al. This in turn makes licensing problematic – you can’t license or monetize exclusive rights you don’t legally have. Another risk is that anyone could copy or reuse the AI-generated content without permission, since copyright wouldn’t restrain them.

Indonesia’s Directorate General of Intellectual Property (DJKI) has acknowledged these issues. Officials urge creators to maintain a “human touch” in AI-assisted works to ensure they remain protected. “Only works that have human creative value can be fully protected by copyright,” explained Ignatius Mangantar, Director of Copyright at DJKI​. If an AI like ChatGPT writes an article or a Midjourney-like AI generates an image entirely on its own, the result “might not meet the criteria for copyright protection due to lack of human originality”​. The implication: authors and artists who use AI tools should add enough of their own creative input – in selecting prompts, curating or editing outputs – so that the final work can be deemed a human creation. DJKI advises creators to understand the limits of AI use in their creative process and ensure the human element remains central, so their works “remain recognized and protected under the law.

Key copyright challenges for AI-generated works include:

• Authorship – Indonesian law recognizes only human or legal persons as authors. An AI cannot hold this status, meaning a purely AI-produced work has no legally recognized author.
• Originality – Works must reflect human creativity and skill​. Content created by AI algorithms may lack the “personal” originality the law demands, unless a human significantly shaped the output.
• Ownership & Rights – Without an author, there is no automatic owner of rights. The work could be effectively unowned (or owned by no one), leaving it free for others to use​. Companies or individuals using AI outputs might find they cannot enforce any IP rights over them.
• Licensing & Contracts – Licensing AI-generated content is legally uncertain if copyright subsistence itself is in doubt. Any license grant could be meaningless if the underlying work isn’t protected. Parties need to carefully structure agreements (and perhaps rely on trade secrets or contracts) when commissioning AI-created works, given the lack of clear IP ownership.

Patents and AI-Generated Inventions: Who Can Be an Inventor? In the patent realm, generative AI is now contributing to inventions – from suggesting new chemical compounds to optimizing engineering designs. However, patent law has a similar human-centric foundation. Indonesia’s Patent Law (Law No. 13 of 2016, recently amended by Law No. 65 of 2024) defines an invention essentially as a product of an inventor’s idea and ingenuity. It implicitly presumes a human inventor behind every invention. While the law does not yet explicitly mention AI, Indonesian IP officials have stated that an AI cannot be named as an “inventor” or “designer” under current statutes. “Indonesia’s Patent and Copyright laws still limit ‘Inventor’ and ‘Author’ only to humans, not AI,” clarified Yasmon, Director of Patents at DJKI, in early 2024​. In other words, if an invention were autonomously generated by an AI with no human contribution, there is presently no legal mechanism in Indonesia to recognize that AI as the inventor on a patent application.

This means that for any patent application arising from AI-assisted innovation, a human must be named as the inventor – typically the person (or team) who conceived the idea or who guided the AI in producing the invention. If no natural person can legitimately claim to have contributed to the inventive concept, the patent would likely be refused for lack of a valid inventor. The Indonesian Patent Office (DJKI) would not grant a patent where the “inventor” field is left to a machine. This aligns with the principle seen globally: patent systems hinge on human inventorship as a core requirement for validity.

However, drawing the line between AI assistance and human inventiveness can be tricky. Consider an engineer who uses an AI algorithm to autonomously design a new product. If the AI’s output is essentially the entire inventive step, what did the human do? Around the world, patent offices and courts are grappling with exactly this scenario. A famous example is the DABUS AI invention case. Dr. Stephen Thaler listed his AI system “DABUS” as the sole inventor on patent filings for a food container and a flashing light device in multiple jurisdictions. The outcome? Virtually all patent authorities said an AI cannot be an inventor. For instance, after a brief initial win, Australian courts overturned a decision and ruled that only a human can be an inventor – and invention “ultimately arises from the mind of a natural person”, so a machine cannot be listed​. The Australian Patent Office and courts indicated that if AI is heavily involved, one should name the human who developed or used the AI, so long as that person made an inventive contribution​. Similarly, U.S. and European patent offices have rejected patents with non-human inventors, and the same would hold in Indonesia under current law.

Key patent challenges for AI-generated inventions include:

• Inventorship – Patent law requires a named inventor (or inventors) who are legal persons. Indonesian law does not recognize AI as an inventor, so any patentable invention must be attributed to one or more humans​. Failing to identify a human inventor means no patent can be granted
• Human Contribution – Determining how much human input is enough to qualify someone as an inventor is a grey area. If an AI generates an invention idea with minimal human direction, there may be debates over whether any person truly “invented” it. Best practice in such cases (per global guidance) is to identify those who conceived the core idea or who built and trained the AI and document their contributions​.
• Patentability – Aside from inventorship, AI-generated inventions still must meet patentability criteria (novelty, inventive step, industrial applicability). The new Indonesian Patent Law (2024 amendment) broadened the scope of patentable subject matter (e.g. explicitly allowing computer programs with technical effects to be patented)​. But it did not change the inventorship requirement – a reminder that legal novelty isn’t enough; a human inventor is still mandatory.
• Ownership of Inventions – Typically, patent rights initially belong to the inventor or their employer. If an AI were ever credited (in a future legal regime) as inventor, issues would arise over who owns the patent (perhaps the AI’s owner or operator). For now, since an AI cannot hold rights, ownership defaults to the human applicant who files for the patent. Companies using AI to invent should ensure employment or contractual agreements made clear that human team members will be named as inventors and the company will own the resulting patents.

International Approaches: How Other Jurisdictions Handle AI-Generated IP Indonesia is not alone in facing these dilemmas – governments worldwide are assessing how to adapt IP law for AI. Approaches vary:

• Australia: Australia grabbed headlines with the Thaler “DABUS” case on AI inventorship. In 2021, a federal court initially allowed an AI to be named as inventor, a world-first. But in 2022 the Full Federal Court unanimously reversed that decision, confirming that under Australian law an inventor must be human​. The court emphasized that patentable inventions must originate from a natural person’s mind, though it noted the AI’s developer or user could potentially be credited if they contributed to the inventive concept. On the copyright side, Australia has so far taken a traditional stance similar to Indonesia – copyright subsists only in works with human authorship (there’s no special statute for AI-generated works yet, as the UK has). Australian authorities continue to review their IP laws, but no explicit protection for AI-generated creations exists at this time.
• United Kingdom: The UK is often cited for a more AI-friendly twist in copyright. UK copyright law explicitly provides that for “computer-generated” works (where no human author is identifiable), the author is deemed to be the person who undertook the arrangements necessary for the creation​. This means a purely AI-generated artwork or literary work can get copyright in the UK, with the rights assigned to the programmer or user who set up the AI’s output. (Notably, the duration of protection for such works in the UK is shorter – 50 years instead of the usual 70 after death of author.) This UK approach is unique and often held up as a possible model to recognise AI-assisted works while still tying them to a human proxy. In patent law, however, the UK aligns with others in not allowing AI inventors – the UK Intellectual Property Office and courts also refused Dr. Thaler’s DABUS applications, insisting on human inventorship.
• Russia: Russia has signalled an even more radical approach. In late 2020, the Russian government drafted a law to grant copyright for AI-created content to the “owner” of the AI​. The proposal covers AI-generated audio, visual, textual and other content, and aims to clarify that the AI itself cannot own the copyright, but its owner (such as the company or person running the AI) would hold the rights​. This would be a novel solution, essentially assigning authorship by statute to a human affiliated with the AI. Russian experts debated whether the AI’s user (who prompts it) rather than the software owner should be the rightful rightholder, likening it to how a digital artist, not the software maker, owns artwork created in Photoshop​. As of now, it’s unclear if this draft has become law, but it shows one way lawmakers are exploring to explicitly integrate AI into IP regimes – a stark contrast to the wait-and-see approach elsewhere.
• United States and Others: The U.S. has taken a firm stand that pure AI-generated works lack the human authorship required for copyright​. The U.S. The Copyright Office and courts have denied registration to AI-created images and writings, though they allowed that human-authored elements (like the selection of prompts or any editing) can be protected​.

On patents, U.S. law similarly demands human inventors and the USPTO has rejected non-human inventor filings. The European Patent Office and most jurisdictions follow the same human-only inventorship rule. International bodies like the World Intellectual Property Organization (WIPO) have been holding discussions and issuing papers on AI and IP since 2019, but no global treaty or uniform standard has emerged yet. This means each country is currently feeling its way through the generative AI era with potentially differing solutions.

Is Indonesian Law Ready? – Ongoing Reforms and Future Outlook Indonesian law, as it stands, was not written with creative machines in mind. Recognising this, Indonesian authorities have started to discuss reforms to ensure the IP system remains effective in the AI age. Issues of “ownership of creations due to AI technology” have been explicitly flagged by officials as a new challenge requiring attention​. The Director General of IP, Razilu, noted in late 2024 that the digital era – including the rise of AI, blockchain, and IoT – “demands flexible legal rules to accommodate ever-evolving innovation.” He emphasized the importance of strengthening regulations to protect digital works in an adaptive way​. In fact, the government has made IP in the digital era a focus for 2025, launching educational programs and pushing awareness about protecting creative and innovative works in a technologically advanced environment.

On the legislative front, there are signs of movement. DJKI has revealed that it is working on revising the Copyright Law to address the needs of the digital era​. While details are still under wraps, this initiative could potentially clarify the status of AI-assisted works – for example, by defining how much human input is required for copyright, or by introducing provisions (perhaps inspired by the UK model) to assign authorship in computer-generated works. Any amendment would likely aim to “ensure creators get sufficient protection while using technology responsibly, so their copyrights remain safeguarded,” as noted by DJKI’s Ignatius Mangantar​. Stakeholders in Indonesia’s creative industries are eager to see if the forthcoming changes will explicitly tackle generative AI, providing much-needed legal certainty.

In the patent sphere, Indonesia’s recent amendments in 2024 modernised various aspects of patent law (such as allowing a grace period and broadening patentable subject matter)​, but did not yet directly address AI inventorship. It remains to be seen if future regulations or guidelines will be issued by DJKI to advise how patent examiners should handle inventions developed with AI assistance. Given global trends, Indonesia may adopt a conservative stance – continuing to require a human inventor on paper, but perhaps issuing guidance for researchers on documenting human contributions when using AI in R&D.

Looking ahead, the legal landscape for AI-generated IP in Indonesia will likely evolve through a combination of law reform, policy guidance, and possibly court decisions if disputes arise. Indonesian policymakers are balancing the need to reward human creativity and innovation (the core rationale of IP rights) against the goal of fostering technological advancement. Clear rules would help – whether that’s affirming that purely AI-made works are unprotected (thus pushing creators to involve humans), or carving out new protections to encourage development of AI creativity. As Indonesia refines its approach, it will draw lessons from the international experiences discussed above, while crafting rules suited to its own legal principles and economic interests.

Conclusion: Navigating the AI-IP Frontier The advent of generative AI is reshaping creativity and invention, and Indonesia’s IP laws are at a crossroads. Is Indonesian law ready for the generative AI era? Not entirely – but it is catching up. In the meantime, businesses and creators using AI should proceed with caution. Keep a close eye on legal developments and engage with IP advisors to understand how emerging rules might affect your AI-driven projects. Before commercialising content or inventions made with AI assistance, it’s wise to consult legal experts and consider protective measures (contracts, disclaimers, or alternative IP strategies) to mitigate risks. By staying informed and proactive, innovators can embrace AI’s benefits while safeguarding their intellectual property interests. The era of “AI vs. IP” is here, and those who prepare will be best positioned to thrive as the laws evolve.