Smarter, Faster, Greener: AI meets natural and sustainable beauty

Emergence

Although artificial intelligence (AI) has been the subject of research and development (R&D) for over 70 years, the modern era of generative AI tools has captured the public attention and seen a steep rise in company applications. In cosmetics, recent years have seen a remarkable expansion of interest and investment in AI, which can be applied across the value chain: from supply chains to new product development (NPD), regulatory challenges, and consumer impact.

Broadly speaking, AI is a non-human technology that learns from information provided to it to recognise patterns and complete tasks. There are various AI models that can be created using a combination of algorithms and data. Training the model is essential to analyse datasets, find patters, make predictions, and generate content. For practical use, understanding the distinctions between AI, machine learning, and deep learning can help users choose the right technology for each application.

Nevertheless, AI is not without human intervention, which is still needed to understand data and perform tasks outside the scope of its training. Data is essential for predictions and decisions: the more data an AI model has, the more accurate it becomes. Consequently, access to data - and importantly how to treat data gaps necessary for predictions - remains a challenge. Finally, instances where protected data subject to intellectual property (IP), such as copyright or patents, are used to train AI increasingly test traditional legal frameworks.

Natural Needs, AI support

By focusing almost exclusively on natural substances, the natural (and organic) cosmetic sector requires information to guide choices of ‘green’ alternatives to conventional (fossil fuel-based) ingredients, as well as identify key factors influencing sustainability across a product’s life cycle: concept and (eco)design, sourcing, manufacture and packaging, distribution, consumer use, and the post-consumer phase.

With the integration of AI into multiple tools, its tangible benefits need to be clear to dispel this technology from a buzzword to something that has real-world value and offers a competitive advantage. The natural sector already needs the ability to confidently select and distinguish raw material based upon origin (genetic modification, ethical engagement, biodiversity impact), and environmental fate (biodegradability).

Nevertheless, using AI in NPD to screen prototypes and optimise R&D, reduce waste, and evaluate emissions (GHGs) is particularly relevant for naturals, where formulations often combine mixtures natural complex substances (NCSs) as actives for skin conditioning, fragrance blends, or specific functional purposes. Indeed, AI can also support naturals in ingredient discovery, safety assessment, environmental impact, regulatory compliance, market research and claims comparisons. Yet, again, key components to evaluate are algorithmic bias and data quality, especially if dataset gaps are plugged by simulations or extrapolations that can affect the reliability and relevance of the results.

Safety and Claims

Most operators in the natural sector are SMEs, making democratisation of AI access a key factor for the widespread application of this technology to the sector. Early adoption of AI is often not possible for smaller operators due to cost, limiting their ability to capitalise on the latest advances or access necessary data.

First and foremost, all cosmetics must be safe. AI can accelerate safety assessments (predictive toxicology), reduce animal testing – for example, via recent irritation prediction model, and evaluate product compliance (safety reports, product information files) within evolving international regulatory frameworks. However, human oversight remains essential to comply with the law. For example, in the EU, under Regulation (EC) No. 1223/2009, a qualified safety assessor must sign the Cosmetic Product Safety Report (CPSR), and a Responsible Person (RP) ensures compliance before a product is placed on the market, irrespective of whether AI was used.

Interconnectivity between AI use, human oversight and legal compliance is not limited to the cosmetic product regulation. Green Claims are essential to natural cosmetics, and AI introduces new risks under the forthcoming application of Directive (EU) 2024/825. For instance, AI may be used to generate the data that is used to substantiate the product claim, which accordingly means if the data is misleading, unverifiable or plain incorrect, then the claim becomes unsubstantiated. Due to the risk of algorithmic bias, AI could potentially generate data that is less objective – possibly through overreliance on information of low quality or low credibility. At the same time, the omnipresence of AI can increase scrutiny from regulators, NGOs, competitors and consumers. Evidently, the ability to rapidly review product claims to flag inconsistencies between official statements, reports and product claims, detect non-compliance, or flag potential risk is therefore crucial.

Sustainability: Reward versus Risk

From an ethical perspective, the natural sector is closely tied to sustainability given the fundamental reliance on renewable raw materials and the principle of defossilising processes. The rewards of AI - rapidly identifying more sustainable pathways and developing innovation products better for the environmental - has a trade-off. Just by its use, there is an ‘AI buy-in’ price to pay. Whether an eventual ‘AI bubble’ or not, high-level investment in data centres requires large land areas, coupled with massive power capacities and server liquid cooling with high grade water, all of which have environmental impacts. The energy consumption alone is high with typical facilities in the 5-10 MW range, although hyperscale centres operate at greater than 100 MW. Equally, a centre’s energy needs are often not supported by renewables, meaning reliance on fossil fuels remains and may grow as demand for AI does.

Conclusion

As an innovation tool, AI has the potential to enable, influence and reshape the cosmetics landscape towards more data-driven, rapid, safe, and sustainable product development. Yet there is a clear need to balance these benefits with transparency in data use and its limitations, algorithmic governance, and the intrinsic resource impact of growing AI investment.