Panel discussion on Beauty Tech: Where Science, AI, and Personalization Meet

 

Predictive skin diagnostics, virtual try-ons, real-time personalized formulations, connected devices seamlessly integrated into daily routines: Beauty Tech is no longer a distant horizon, it is an accelerating present. To assess what is truly at stake, 13 international experts drawn from clinical research, formulation science, regulatory consulting, active ingredient design, and trend analysis offer a nuanced picture of Beauty Tech in 2025–2026: ambitious in its scope yet demanding in its conditions for success. Data quality, consumer trust, and regulatory vigilance remain chronically underestimated.

 

Cross-analysis of contributions reveals several major dynamics at work. The first, and most pervasive, is the rise of AI-powered skin and hair diagnostics. Optical technologies, automated trichoscopy, multi-spectral imaging now enable precise quantification of biomarkers: hair density, pigmentation, and skin barrier integrity. Multi-omics analysis (genomics, proteomics, metabolomics) is emerging as the next analytical frontier. The concept of a skin “Digital Twin” a digital counterpart integrating biological data with environmental factors, is taking shape as the sector’s defining horizon. The second axis is augmented reality. Long confined to entertainment, AR try-on must now demonstrate genuine decision-making value, measured not by engagement metrics, but by its impact on conversion rates and product return reduction. The critical distinction lies between visualization (showing a product) and simulation (reproducing its actual behavior on a specific skin type, whether in terms of color or texture). Personalization forms the third element of this new technological landscape. Contributors carefully distinguish two levels of maturity:

• surface personalization (quizzes, skin-type profiling) and precision personalization, which combines biometrological, behavioral, and environmental data in real time.
• True personalization demands data continuity across the entire value chain from consumer insight through to industrialisable formulation.

The fourth axis is the integration of connected devices, LED masks, IoT-enabled applicators, smart packaging, into holistic routines where topical products and tech tools reinforce one another. Systems such as Amorepacific’s Skinsight™ (a CES 2026 honoree), L’Oréal’s Cell BioPrint (CES 2025), and NuraLogix’s Anura MagicMirror embody this shift towards accessible, predictive, and personalised diagnostics. When a product becomes part of a connected routine, packaging itself changes in nature: it must integrate dispensing precision, hygiene, refill logic, and ritual design.

 

If one conclusion runs through the entire corpus, it is this: AI in cosmetics only performs when it rests on reliable, structured, inclusive, and diverse data. The challenge is not algorithmic, it is infrastructural, spanning data management from personalized diagnosis through to targeted product recommendation. The fragmentation of data across consumer applications, R&D systems, and supplier databases remains the primary barrier to AI’s real impact across the beauty value chain. On the formulation side, R&D agility is the true bottleneck of personalization: consumer insights are insufficient if they cannot be translated into industrialisable prototypes in real time, within the regulatory constraints of each target market. A critical tension emerges when 80% of consumers are open to personalization, yet two thirds have already experienced inaccurate or intrusive interactions that have eroded that trust. In a domain as intimate as skin care, the margin for error is slim. Claims of systematic AI superiority must be tempered by clinical evidence: where a trained human expert demonstrates greater sensitivity than an algorithm at early measurement timepoints, this must be acknowledged and integrated into the process. Human–AI complementarity, not substitution, stands as the scientifically sound posture.

 

Despite the diversity of profiles and sectors represented, several convictions run through all contributions without exception. First, AI represents a structural transformation, not a passing trend: all contributors agree on the irreversibility of this integration into the cosmetics industry. Then, data, its quality, diversity, and structuration, is the true limiting factor, far more so than algorithmic sophistication. Finally, consumer trust is a non-negotiable prerequisite for large-scale adoption: transparency regarding data use, rigorous clinical validation, and the interpretability of recommendations are indispensable. South Korea is unanimously recognized as the global laboratory of Beauty Tech. K-Beauty 2.0 is less an aesthetic trend than a structural shift, integrating speed, iteration, technology, and ritual into a model capable of translating a consumer insight into an industrialisable prototype in under six months. This leadership rests on solid foundations: semiconductor dominance, a national AI strategy, and a unique symbiosis between deeply anchored daily rituals and cultural openness to digital tools. Is K-Beauty a source of inspiration or a transposable model? Some experts advocate its direct replicability; others argue that what matters most is not geography, but the density and continuity of feedback loops between consumers and R&D. The lesson of K-Beauty is less a model to copy than a mindset to adopt.

 

Tool or paradigm? Some experts position AI primarily as a process optimization tool (molecular design, clinical analysis). Others see it as a paradigm shift requiring a wholesale rethinking of data architecture and business models across the sector. Most industry contributions adopt a prospective, enthusiastic tone. Some experts introduce a note of institutional caution: under the combined effect of the GDPR and the EU AI Act, certain cosmetic applications, diagnostics, biased recommendation engines, behavioral analytics, may fall into the “high-risk” category. Regulatory analysis at the design stage is not an option; it is a requirement. This reflects both a well-founded industrial optimism and an indispensable regulatory prudence.

 

Although frequently absent from marketing discourse, meaningful limitations do exist. AI systems perform well for profiles well represented in training datasets, but struggle with edge cases: specific sensitivities, rare combinations, under-represented regional needs. “Universal” personalization remains a horizon, not yet a fully achieved reality. Ultimately, the most promising Beauty Tech is not the one that accumulates the greatest number of features. It is the one that solves a genuine problem, grounded in reliable data, rigorous validation, and an experience worthy of consumer trust. True personalization demands data continuity from upstream to downstream, from consumer insight to industrialisable formulation. This is precisely what South Korea is demonstrating at industrial scale, and what the broader beauty world is now invited to integrate into its own practices, in order to keep innovating and meeting the evolving expectations of its consumers.

 

Ivona Ferić
Content Manager, Arbelle

–
Davide Musardo
Claims & Efficacy Team Lead, Biorius

–
Cécile Guyot
Communication Manager, Coptis

–
Marie Magnan
Regulatory Affairs Manager, COSMED,
the French cosmetic Association for SMEs

–
Yann Chilvers
Founder & Co-CEO, Covalo

–
Alexander Kwapis
Global Head of Innovation, R&D, and Engineering FusionPKG an Aptar Beauty Company

–
Mallory Huron
Director of Beauty + Wellness, Future Snoops

–
Bum chun Lee
CEO of Huenskin Co. Ltd.

–
Raya Khanin
Chief Scientific Officer and Co-founder, Lifenome

–
Andrea Esplugas
Marketing Manager, Lipotrue

–
Pascale Gauthier
Pharmacist PhD, Charge of courses, Lecturer, Auvergne University, Faculty of Pharmacy

–
Seongmin(Mike) Sohn
CEO & Principal Consultant, Wynngate Korea Co., Ltd.

–
Laura Cabrera
Haircare technical director, Zurko Reasearch