Rethinking patentability in the age of AI and biotechnology: a Peruvian perspective

Maritza Reátegui
Partner, RODRIGO, Lima, Perú
mreategui@estudiorodrigo.com

Daniela Supo
Senior Associate, RODRIGO, Lima, Perú
dsupo@estudiorodrigo.com

Artificial intelligence (AI) is transforming biotechnology at extraordinary speed. In drug discovery, genomic analysis, biomarker identification, protein engineering, diagnostics, and precision medicine, AI no longer operates merely as a support tool. It increasingly shapes how research questions are framed, how biological candidates are selected, and how experimental results are interpreted. This convergence raises a difficult question for patent law: should patentability standards be reconsidered when innovation emerges from the interaction between computational systems and biological research? From the perspective of Peru, the answer is yes, but not because a wholly new patent regime is required. Rather, what must be reconsidered is the interpretation and application of existing standards under Andean Decision 486, Legislative Decree No 1075, and the Andean Manual for Patent Examination (which expressly recognises the contemporary importance of biotechnology, computer-implemented inventions, and AI in Andean patent practice).

Peru’s patent system is rooted in Andean Decision 486, which establishes that patents are available for inventions, whether product or process, in all fields of technology, provided they are new, involve an inventive step, and are capable of industrial application. Decision 486 also defines what is not considered an invention and what is not patentable. Article 15^[1] excludes, among other things, discoveries, scientific theories, mathematical methods, biological material as found in nature, and computer programs ‘as such’. Article 20^[2] further excludes plants, animals, essentially biological processes for the production of plants or animals, and therapeutic, surgical, and diagnostic methods applied to humans or animals. Any Peruvian analysis must begin by asking not whether AI was used, but whether the claimed subject matter constitutes a patentable technical solution under this framework.

That distinction is especially important because AI-driven biotechnology inventions are often hybrid in nature. They may involve a machine-learning model, a biological dataset, a screening method, a predicted sequence or compound, and subsequent wet-lab validation. Some of those elements may be non-patentable if claimed in the abstract. A predictive model, standing alone, may fall within the exclusions for mathematical methods or computer programs as such. Likewise, naturally occurring biological material remains excluded merely because it has been identified, isolated, or computationally characterised. By contrast, a claim directed to a concrete technical application, such as a reproducible method, an engineered biological construct, a non-natural product, or a technical process that uses AI as part of its implementation, may still fall within patentable subject matter. The Andean Manual for Patent Examination is especially useful here because it does not treat software, AI, and biotechnology as isolated categories. It includes dedicated sections on artificial intelligence and machine learning, on computer-implemented inventions, and on ‘hybrid cases with chemistry and biotechnology’, thus signaling that modern examination in the Andean system must address these overlaps in a structured way.

The first patentability standard that requires closer scrutiny is inventorship. Decision 486 clearly presupposes a human inventor: Article 24^[3] states that the inventor has the right to be mentioned as such in the patent, and Article 27 requires identification of the inventor in the application. The Andean Manual reinforces this point by stating, at the formal examination stage, that the inventor must always be one or more natural persons. Under Peruvian law, then, there is no basis for recognising an AI system as an inventor. But that does not resolve the more difficult issue. The real challenge is determining what level of human contribution is sufficient where AI materially shapes the inventive process. A defensible Peruvian approach would preserve human inventorship while requiring evidence of meaningful human intellectual intervention: identifying the technical problem, selecting the relevant variables and training criteria, interpreting the computational output, and translating it into a reproducible technical solution. In biotechnology, that human contribution remains crucial because the gap between a computational prediction and a patentable invention often lies in experimental design, validation, and technical embodiment.

The second standard that must be reconsidered is inventive step. Article 18^[4] of Decision 486 asks whether the invention would have been obvious to a person ordinarily skilled in the art. The Andean Manual develops this inquiry in detail, including the problem-solution method and technology-specific guidance, including for biotechnology and computer-implemented inventions. This is where AI changes the factual setting. If machine-learning-based screening, sequence comparison, protein modelling, or candidate prioritisation becomes routine in a technical field, then the mere use of AI should not be enough to support inventiveness. Yet the opposite risk also exists: patent offices may overvalue AI-assisted outputs simply because they appear computationally sophisticated. Peru should avoid both errors. The proper inquiry is whether the claimed invention represents a non-obvious technical advance over the prior art, not whether it was produced with powerful software. AI should neither automatically elevate nor automatically negate inventive step. Instead, the examiner should ask whether the claimed result solved a technical problem in a non-evident manner, yielded an unexpected effect, or produced a result that the prior art did not render reasonably predictable. This interpretation is entirely consistent with Decision 486, but it requires greater discipline in examination. 

If inventive step becomes more nuanced in AI-biotechnology, sufficiency of disclosure becomes even more important. Decision 486 already provides a strong framework. Article 28 requires the invention to be disclosed in a manner sufficiently clear and complete for a skilled person to understand and carry it out. Article 29^[5] requires a deposit where the invention refers to biological material that cannot otherwise be adequately described. Article 30^[6] adds that claims must be clear, concise, and fully supported by the description. The Andean Manual gives practical content to these provisions. It explains that the description must allow the skilled person to understand the technical problem and solution and to reproduce the invention without inventive effort; it also states that where the scope of the claims is excessive relative to the disclosed embodiments, sufficiency is lacking. The domestic complementary framework points in the same direction: Legislative Decree No 1075 provides that disclosure is sufficiently clear and complete when a person skilled in the art does not need further experimentation in order to carry out the invention as of the filing date, and when the description reasonably shows that the applicant was in possession of the invention at that time.

These rules have direct consequences for AI-driven biotechnology. An applicant should not obtain broad protection over a biologically significant result if that result depends on opaque model architecture, undisclosed input parameters, hidden training assumptions, or insufficient validation. That does not necessarily mean that source code must be disclosed in every case. But where AI is central to identifying a sequence, compound, biomarker, or biological pathway, the application should disclose enough technical information to make the claimed contribution reproducible. In biotechnology, weak disclosure is particularly problematic because overbroad claims can block follow-on research while contributing little to public knowledge. Peru, therefore, does not need a new rule, it needs a stricter application of existing sufficiency, support, and clarity requirements. The Andean Manual is especially significant because it expressly connects insufficient disclosure to excessive claim breadth and states that new examples cannot later be added to cure an original lack of sufficiency, since that would amount to an impermissible extension beyond the initial disclosure.

A Peruvian perspective also requires taking biodiversity governance seriously. The relevant framework extends beyond Andean Decision 486^[7] and includes the Andean regime on access to genetic resources, particularly Decision 391. Read together, these instruments make clear that patent protection in the Andean system is not a purely technical exercise, but one that must remain attentive to lawful access to genetic resources, the protection of biological and genetic heritage, and the collective knowledge of indigenous, Afro-descendant, and local communities. Article 26 further requires, where applicable, a copy of the access contract when the invention was obtained or developed from genetic resources, and a document evidencing authorisation to use traditional knowledge when relevant. The Andean Manual confirms that, during formal examination, the office must verify the existence of the contract or other document proving lawful access to those resources or knowledge, and it devotes an entire annex to genetic resources and access contracts. Legislative Decree No 1075 adds further force by providing sanctions for non-compliance with the rules applicable to genetic resources and traditional knowledge.^[8] This dimension is especially relevant in AI-driven biotechnology because AI can accelerate the extraction of value from biological information, including information derived from biodiversity-rich countries such as Peru. For Peru, then, reconsidering patentability standards cannot mean loosening scrutiny. On the contrary, it means ensuring that AI does not become a mechanism for repackaging biological appropriation as algorithmic innovation.

In this sense, Peru should adopt a middle path. It should not create a sui generis category for AI-assisted biotechnology inventions, but neither should it mechanically apply older categories without regard to technological reality. The better position is to preserve the current legal architecture while interpreting it more rigorously. Inventorship should remain human. Inventive step should focus on genuine technical non-obviousness, not on the mere use of AI. Sufficiency of disclosure should be applied more strictly where AI affects reproducibility and claim breadth. And biodiversity-related safeguards must remain central wherever the invention relies on genetic resources or associated traditional knowledge. If approached in this way, Peru can modernise patent examination without weakening the public-interest structure embedded in Andean patent law.

Ultimately, the central challenge is not whether AI transforms patent law beyond recognition, but whether the patent system can continue to distinguish genuine technical invention from the mere acceleration of research through computational tools. From a Peruvian perspective, that distinction remains both possible and desirable within the existing legal framework. The task, therefore, is not to create a new patent regime, but to apply the current one with greater rigor, technical sophistication, and sensitivity to the specific risks posed by AI-driven biotechnology. Properly understood, rethinking patentability standards in Peru means demanding a more disciplined and exacting examination, one capable of rewarding true innovation without diluting the legal limits that protect the public domain, biodiversity, and traditional knowledge.

Notes

^[1] The following shall not be considered inventions:

1. discoveries, scientific theories and mathematical methods;
2. the entirety or part of living beings as encountered in nature, natural biological processes, biological material existing in nature or which may be isolated, including the genome or germ plasm of any natural living being;
3. literary and artistic works or any other work protected by copyright;
4. plans, rules and methods for the pursuit of intellectual activities, the playing of games or the conduct of economic and business activities;
5. computer programs or software as such; and
6. methods of presenting information.

^[2] The following shall not be patentable:

1. inventions the commercial exploitation of which on the territory of the member country concerned has necessarily to be prohibited in order to protect law and order or morality. To that end the commercial exploitation of an invention shall not be considered contrary to law and order or morality solely owing to the existence of a legal or administrative provision that prohibits or regulates such exploitation;
2. inventions the commercial exploitation of which in the member country concerned has necessarily to be prohibited in order to protect the health or life of persons or animals, or to preserve plants or the environment. To that end the commercial exploitation of an invention shall not be considered contrary to the health or life of persons or animals or liable to prejudice the conservation of plants or the environment solely on account of the existence of a legal or administrative provision that prohibits or regulates such exploitation;
3. plants, animals and essentially biological processes for the production of plants or animals that are not non-biological or microbiological processes;
4. therapeutic or surgical methods for the treatment of human beings or animals, and also diagnostic methods applied to human beings or animals.

^[3] The inventor shall have the right to be mentioned as such in the patent and may likewise object to being so mentioned.

^[4] An invention shall be regarded as involving an inventive step if, for a person in the trade with average skills in the technical field concerned, the said invention is neither obvious nor obviously derived from the state of the art.

^[5] ‘Where the invention relates to a product or process involving biological material, and the invention cannot be described in such a way as to be understood and carried out by a person skilled in the technical field, the description shall be completed with a deposit of the said material.

The deposit shall be made not later than on the filing date of the application in the member country or, where applicable, on the filing date of the application the priority of which is claimed. Deposits effected with an international depositary authority recognized under the Budapest Treaty on the International Recognition of the Deposit of Microorganisms for the Purposes of Patent Procedure of 1977, or with another institution recognized by the national office competent in that connection, shall be considered valid. In such a case the description shall state the name and address of the depositary institution, the date of deposit and the deposit number assigned to it by the institution.

The deposit of biological material shall be valid for the purposes of the grant of a patent only if it is made in such a way that any interested party may obtain samples of the material not later than on the expiry of the period provided for in Article 40.’

^[6] The claims shall define the subject matter to be protected by the patent. They shall be clear and concise and entirely supported by the description. The claims may be either independent or dependent. A claim shall be independent where it defines the subject matter to be protected without reference to another, earlier claim. A claim shall be dependent where it defines the material to be protected by reference to an earlier claim. A claim that refers to two or more earlier claims shall be considered a multidependent claim.

^[7] Member countries shall ensure that the protection conferred on the various forms of industrial property shall be granted in such a way as to safeguard and respect their biological and genetic heritage and also the traditional knowledge of their indigenous Afro-American or local communities. By virtue of the foregoing, the grant of patents relating to inventions developed on the basis of material derived from that heritage or knowledge shall be subject to that material having been acquired in accordance with international, community and national legal provisions.

Member countries shall recognize the right and competence of indigenous Afro- American or local communities to decide on matters pertaining to their collective knowledge.

The provisions of this Decision shall apply and be interpreted in such a way that they do not contravene those laid down in Decision 391 in its currently amended form.

^[8] Article 120-A: Non-compliance with the rules applicable to genetic resources and traditional knowledge: Failure by a patent applicant to comply with the requirement to submit the agreement referred to in Article 26(h) and (i) of Decision 486 of the Andean Community Commission, as further developed in Articles 20 and 21 of the Regulation on Access to Genetic Resources, shall give rise to one or more of the following sanctions, unless the applicant withdraws from the patent grant proceeding or provides a satisfactory explanation that the invention does not use such traditional knowledge or genetic resource of which Peru is the country of origin:

1. A fine of up to 1,000 UITs.
2. Compensation.
3. Fair and equitable benefit-sharing, including the sharing of royalties and/or other monetary or non-monetary measures.
4. Technology transfer and capacity building.
5. Use authorizations.

Whenever compulsory patent licenses are involved, Articles 61 to 69 of Decision 486 of the Andean Community Commission and Article 40 of this Legislative Decree shall apply.