The digital era has brought immense technological change in all aspects of life and society with the rise of the internet, mobile devices, cloud computing and more. GenAI has now emerged to shape code development, art, music and much more as it becomes ubiquitous in daily life. However, an even bigger transformation may be on the horizon - the advent of generative artificial intelligence applied to biology and biotechnology. Just as tools like ChatGPT can generate human-like text based on data training, genAI technologies could eventually design new proteins, metabolic pathways, x organisms and biological processes at a vastly accelerated pace. We may well be on the cusp of a biotech "ChatGPT moment" that will reshape industries and economies. As it stands today, many companies are already embracing genAI to help speed up processes in research and development (R&D).
The Potential of GenAI in Biotech
The potential applications of genAI in biotech are significant. Researchers are already using AI to predict protein structures, model disease progression, and test theoretical biological experiments in silico for example before running costly wet lab trials1 However, this is just the start of what might be possible. Advanced genAI could one day:
- Design optimised enzymes for industrial processes or bioremediation
- Create personalised cell therapies by engineering immune cells to target cancers
- Develop drought/heat resistant crops to boost food security
- Evolve microbes to produce biofuels, biomaterials, pharmaceuticals and more
- Predict disease risk, progression and optimal treatment on an individualised basis
- By augmenting human intelligence, genAI may help overcome bottlenecks in R&D, accelerating innovation cycles for everything from sustainable materials to regenerative medicine. The economic impacts could be staggering, with the bioeconomy already generating 5 percent of U.S. GDP2.
Challenges and Risks
The power of these technologies also raises challenges around ethics, security, intellectual property and more. Just as generative text or photo/art/music AI raises issues around misinformation and copyright, genAI biotech applications demand robust governance, perhaps even more so.
Biosecurity is a prime concern - theoretically, bad actors could use genAI to design bioweapons or other threats more easily. While the actual creation of engineered pathogens still requires physical resources, safeguards are needed. Approaches like biosecurity screening of AI models and tighter controls on ordering gene sequences and biomolecules may be required.
There are also difficult questions around intellectual property and data rights for AI-generated biological assets like organisms or molecular therapies.
If genAI designs a novel enzyme, cell line or microbe, who owns the intellectual property (IP) - the AI developers, companies who provided training data, or neither?
In fact, questions are now being raised around the training data itself which may be subject to copyright or other intellectual property protections. New legal and regulatory frameworks are going to be required.
Promoting Responsible Innovation
To ensure the benefits of genAI outweigh the risks, a focus on responsible innovation and ethical, human-centred design is critical. This means:
- Proactive governance - Policymakers should start consulting now with scientists, industry, ethicists and security experts to craft appropriate guidelines and regulations before genAI biotech goes mainstream. Public-private collaboration is key.
- Risk assessment - Rigorous risk analysis on potential misuse cases is required to identify vulnerabilities and safeguards. This includes evaluating genAI's capabilities around areas like bioweapons design or intellectual property infringement.
- Ethical training - AI systems must be carefully trained on the principles of bioethics, biosafety and biosecurity to align with human values. Lessons can be drawn from initiatives around constitutional AI and medical ethics.
- Access controls - While genetic engineering tools are increasingly accessible, genAI infrastructure may require tighter access controls. This could include identity verification, institutional credentialing and monitoring of associated physical resources.
- Global coordination - GenAI is a global issue requiring international cooperation and norm-setting around development, use and proliferation of these capabilities. No single country can or should do it alone.
Considerations for the Business and Tax Landscape
For companies working in biotech and biomedicine, the genAI wave presents immense opportunities but also a series of strategic considerations:
- Intellectual Property - As noted earlier, the IP situation with AI-generated biological materials and data is grey in a lot of cases. Companies will need to closely track legal developments and case law around patent eligibility, data rights, and AI inventorship. Careful IP management strategies will be essential.
- R&D Location and funding - Where a company locates its R&D activities and how those activities are funded can significantly impact its tax and transfer pricing profile and ability to access R&D incentives.Many countries like Australia, Singapore, the US and the UK offer generous R&D tax credits, patent box, and other grants and incentives for biotech/life sciences innovation carried out domestically. Global organisations will need tax-efficient IP and operational structures and will need to be considerate of global taxation rules like BEPS 2.0 and Pillar Two.
- Transfer Pricing - As AI-powered biomolecular assets get commercialised across borders, transfer pricing will be critical to properly reward value drivers and allocate tax obligations. Companies must implement robust transfer pricing models and documentation.
- Exit Taxation - Monetisation events like selling biotech subsidiaries could trigger exit taxes depending on the jurisdictions involved. Advanced planning is required to understand and mitigate these costs which can significantly affect anticipated return on investments (ROIs).
The genAI biotech revolution offers immense commercial opportunities but also tax/operational complexities that must be strategically managed. Businesses will require sophisticated modelling to maximise value, mitigate risks and manage costs.
A Profound Transformation and Considerations for Cross-Border Business Models
In the coming decade, genAI may transform the $4 trillion bioeconomy just as machine learning disrupted sectors like advertising, finance, and logistics. However, capturing benefits will require foresight.
Business leaders should start reviewing their innovation pipelines, operational footprints, and data strategies now to capitalise on the genAI wave. Policymakers must be proactive in developing robust governance frameworks that promote responsible development while mitigating risks.
Those who are agile in adopting genAI capabilities while embedding ethical principles could reshape entire industries. However, realising that promise hinges on our ability to align technological progress with human values and public good. The genAI biotech revolution is coming and the time to prepare is now.
Biotech and life sciences companies leveraging genAI capabilities need to consider some issues when looking at cross-border business models:
Intra-group services and IP licensing
As genAI platforms are developed and deployed across a company's global operations, there will likely be intra-group charges for providing access to the AI systems, algorithms, training data, etc. The pricing of these intra-group services and IP licensing arrangements needs to be at arm's length.
Companies will need to analyse the functions, assets and risks involved in developing and maintaining the genAI capabilities. This functional analysis will guide the selection and application of the appropriate transfer pricing method (e.g. Cost Plus, TNMM, profit split).
DEMPE functions (development, enhancement, maintenance, protection and exploitation) must align with the locus of the IP in order to maintain the structure of the model. Appropriate substance is critical and ensuring the appropriate rights are captured in legal agreements is critical.
AI-generated intellectual property
When a company's genAI system generates patentable biotech inventions like novel proteins, molecular therapies, etc., there are transfer pricing implications when that IP is exploited across the multinational group.
There needs to be arm's length compensation when the legal owner of the AI-generated IP licenses transfers the rights to other subsidiaries for development or commercialisation. Depending on the circumstances, this could involve cost-contribution arrangements, lump-sum buy-in payments, and royalties priced on arm's length terms.
AI data contributions
The performance and outputs of a genAI system are heavily influenced by the training data. If various subsidiaries contribute proprietary data that enhances the AI capabilities, those data flows may need to be compensated on arm's length terms under transfer pricing principles. This could involve compensating the data-providing entities for something akin to the "manufacture" of a commercialised product generated by the AI using that data.
Cost sharing and co-development
For biotech R&D leveraging genAI across multiple jurisdictions, companies may consider implementing cost sharing or co-development arrangements. This allows the costs and risks of developing the AI-powered innovation to be shared, with resulting revenues or transfers of rights to exploit the IP depending on each entity's contribution.
The key is to ensure costs are shared proportionately to the anticipated benefits, and that any cross-licenses, subsidiary R&D service payments, etc. are priced on arm's length terms per transfer pricing rules.
As genAI accelerates biotech and pharmaceutical innovation across international operations, companies will need to carefully delineate and compensate the functions, assets and risks around developing the AI, generating AI-derived intellectual property, providing training data, mutually sharing costs/risks, and more. Robust transfer pricing models and documentation will be critical to sustaining tax authority scrutiny in this complex area. Given the complex and highly integrated nature of genAI business models, it may be appropriate to consider obtaining bilateral or multilateral advanced pricing agreements with tax authorities. This can help to provide certainty on transfer pricing for a period of time and also help to reduce more generally the potential for tax controversy.
How Can A&M Tax Help?
The A&M tax team has deep tax technical and compliance experience to help navigate these complex new rules, assist with an implementation program and assess how they apply to all types of businesses. Contact us to discuss your particular circumstances further.
Footnotes
1. Read https://www.reuters.com/business/healthcare-pharmaceuticals/sanofi-partners-with-openai-formation-bio-ai-driven-drug-development-2024-05-21/
2. Read https://time.com/6967625/we-need-to-be-ready-for-biotechs-chatgpt-moment/
Originally published 26 June 2024
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.