The beauty industry in 2025 is an example of innovation in action. In the time since our last briefing, the rate of AI adoption has accelerated as businesses seek to keep pace and innovate. For beauty brands, retailers, and consumers, this shift is reshaping how products are created, marketed, and experienced. In the background, other shifts are happening – the EU AI Act is becoming established as the benchmark for AI regulation worldwide, and consumers are becoming increasingly aware of their data rights and concerned about potential misuse. Below, we explore the notable developments that have emerged in the beauty ecosystem since our last analysis, and what those developments might mean for emerging compliance challenges.
Hyper-customisation and “smart” beauty solutions
Personalised product recommendations continue to evolve iteratively. Startups such as Revolve Beauty have pioneered machine learning models that analyse microbiome data to develop tailored skincare regimens. These next-generation tools look beyond a customer's shopping history and skin type, factoring in variables such as environment and diet to provide truly bespoke product recommendations. Some notable newcomers include:
- Shiseido’s 'Optune' platform, which monitors an individual's skin condition and environment (e.g., temperature, humidity) via an app and camera, and automatically dispenses customised skincare lotions that adapt to changing conditions in real-time.
- L'Oréal's recently unveiled 'Perso' system, which uses AI to formulate personalised skincare based on environmental conditions and user preferences. Its capabilities are likely to include the ability to deliver hyper-customisable lipstick shades, too.
- Direct-to-consumer brand Prose has entered the skincare market with AI-powered customisable formulations based on online consultations that also adapt to changing skin conditions.
'Virtual try-ons' evolve
Beyond basic virtual try-on tools, some brands are now advancing into more sophisticated AI-enabled territory. Amorepacific's 'Wanna-Beauty AI', for example, combines face segmentation and large language model (LLM) capability to produce realistic images generated by Stable Diffusion, to offer tailored makeup recommendations and a visual try-on through a chatbot interface with voice recognition.
Multi-platform integration
Increasingly, beauty brands are partnering with technology platforms to expand their virtual offerings across social media and video conferencing software. Maybelline's AI-powered filter for Microsoft Teams is an example of how companies are experimenting with new ways to connect with consumers - particularly those who spend much of their daily life online. Multi-platform solutions, combining a physical thing with AI-powered software, seem to be on the rise. Integrating AI with 'smart' mirrors is emerging, though still in its infancy.
Prototyping
At the recent Cannes Lions International Festival of Creativity, major beauty brands showcased their latest AI applications in product development. The Estée Lauder companies highlighted their innovative approach of combining "math and magic", digitising creative and product archives to generate new campaign and product concepts. Shiseido demonstrated how it analyses user-generated content data with AI to inform product innovation down to the shade level, creating more precise colour matches for consumers. In an effort to understand and stay ahead of emerging consumer needs, some brands are also using AI to generate predictive insights on topics of concern to consumers. By doing so, they are able to rapidly launch targeted solutions and respond to the market faster than ever before.
Branding
Meanwhile, marketing teams are using generative AI models for everything from writing campaign copy to devising influencer marketing strategies. These sophisticated algorithms analyse social media engagement data and consumer sentiment to identify the tone, style, and imagery most likely to resonate with target demographics. Tools like Midjourney and DALL-E are being used to generate visual designs and packaging concepts, while ChatGPT and Jasper help create product descriptions and marketing materials. Often, these tools are used in combination to optimise results, with features like Midjourney's image-to-text capability generating prompts from images that can then be refined using other AI tools. The result is a potentially more agile approach to marketing, but one that calls for scrutiny around IP rights, privilege, bias, and fair representation.
Looking ahead
We anticipate more advanced iterations of custom product formulations, bridging biotech and AI to cater to a broader range of skin tones, allergies, and ethical preferences. The journey toward ever-more specific “made for me” products could help the beauty industry embrace a more inclusive future.
New regulatory Demands: the EU AI Act
Data protection rules have already shaped how beauty and tech intersect. Meanwhile, in terms of AI regulation more generally, the UK's principles-driven approach to AI still centres on safe, transparent, and fair usage. However, the EU AI Act is now being positioned as the roadmap to ensure compliance on a global scale - particularly for those businesses operating in multiple jurisdictions. By embedding “compliance by design,” in line with the principles of the EU AI Act, companies can manage legal risk exposures and build consumer trust from day one.
The EU AI Act classifies AI systems based on the risk they pose to fundamental rights. The most stringent obligations apply to AI systems considered 'high risk'. For these systems, the Act imposes strong obligations around transparency, human oversight, and data governance. Similarly, the Act prohibits AI systems that it deems pose an unacceptable risk – including, for example, AI systems that process biometric data in real time in public spaces.
One thing is apparent when it comes to emerging beauty technologies – most seem to ingest significant amounts of personal and biometric data – first to train and then to function.
The General Data Protection Regulation (GDPR), like the EU AI Act, imposes stringent obligations on companies that process sensitive personal data used to uniquely identify individuals, classified as "special category data". The processing of biometric data, including for facial recognition and skin analysis purposes, may pose significant risks to individuals’ fundamental rights and freedoms if, for instance, it reveals their sensitive health data or ethnicity. Such information, if misused, could lead to discrimination or other adverse outcomes.
To lawfully collect and process special category data, beauty brands must either obtain explicit consent from their consumers or ensure they have another appropriate condition under Article 9 of the GDPR. Additionally, they must be transparent with their users and implement robust safeguards to protect user privacy. Compliance with these requirements is crucial, not only to avoid legal and regulatory risk, but also to maintain consumer trust.
As we look toward a marketplace defined by immersive experiences and unprecedented customisation, the EU AI Act is solidifying its position as the gold standard for responsible AI. To stay ahead of the curve, beauty brands must keep abreast of the regulation before the EU AI Act comes fully into force for all AI systems in August 2026 (certain aspects, including the prohibition on AI posing an unacceptable risk, are already in force). The rewards are considerable for those deploying AI responsibly - boosted brand loyalty, stronger consumer trust, and a dynamic new dimension of customer engagement.
For more information on the EU AI Act and "Compliance by Design", visit our AI Resource Centre. For more information on the IP issues presented by GenAI, please download our GenAI and IP guide, and sign up to receive alerts for our GenAI and IP case tracker.