As artificial intelligence becomes more embedded in beauty workflows, industry leaders foresee the next phase of adoption will extend well beyond personalization and digital try-on.
In this Q&A, we spoke with Anastasia Georgievskaya, CEO and Founder of Haut.AI, Jane Yoo, M.D., Assistant Clinical Professor, Department of Dermatology at Icahn School of Medicine at Mount Sinai, and Wayne Liu, Chief Growth Officer & President of Americas at Perfect Corp.
Here our experts outline how AI is expected to influence formulation science, claims substantiation and strategic investment decisions by 2026, with implications for manufacturers and suppliers across the United States cosmetics and personal care market.
CDU: From your vantage point, how will AI meaningfully change product development and formulation decisions in the beauty industry by mid-2026?
Anastasia Georgievskaya: AI’s most meaningful impact in beauty will move upstream — from personalization and marketing into product development and formulation validation. One of the biggest issues today is that product decisions are often made based on ingredient-level claims, while the finished formulation is rarely evaluated as a whole.
Consumers don’t use ingredients; they use products.
This is where AI can change how decisions are made. In our own work at Haut.AI, we started from clinical software used to measure before-and-after effects in controlled studies, and that experience made it clear how large the gap is between laboratory data and real consumer outcomes.
AI makes it possible to connect ingredient science, clinical insights, and real-world skin data to understand how finished formulations actually perform across different skin types and demographics.
As a result, product teams can iterate formulations based on evidence rather than assumptions. AI will be less about guessing what might work and more about validating what does — earlier, faster, and with greater confidence.
CDU: What AI capabilities do you expect manufacturers and suppliers to realistically adopt at scale this year, and which applications remain overhyped or not yet commercially viable?
Jane Yoo, M.D.: Capabilities that can realistically be adopted at scale include:
- Quality control in manufacturing: Computer vision for fill levels, labeling errors, particulate detection, and batch consistency
- Demand forecasting + inventory optimization: This prevents product discontinuations and reformulation churn.
- Formulation knowledge management: AI systems that allow R&D teams to query internal data (“What happened to viscosity when we swapped X for Y?”).
- Basic predictive toxicology triage: Early screening flags for sensitization risk or ingredient interactions—although this still would need to be tested with trials.
Here’s what is not commercially viable at scale:
- Fully automated “AI makes a new active ingredient” with robust human safety/efficacy data in time for mass beauty launches.
- Claims like “AI proved this reverses aging” without strong clinical validation.
- Biometric personalization requiring extensive consumer data capture (continuous face scanning, real-time physiologic monitoring) as a mainstream model as there are too many privacy (HIPPA)/compliance and bias constraints.
CDU: How do you see AI influencing claims substantiation, safety assessment, and regulatory readiness over the next 18–24 months?
Wayne Liu: This is where AI becomes mission-critical rather than optional. The 2022 Modernization of Cosmetics Regulation Act now requires safety substantiation, serious adverse event reporting within 15 business days, and detailed record-keeping. Manually managing this compliance is challenging for brands operating at scale.
AI platforms can now complete safety assessments that traditionally took months in just minutes. However, the game-changer is the continuous, real-time monitoring that AI enables to identify safety concerns before they escalate.
By late 2026, I expect regulators to increasingly accept AI-generated documentation, but with caveats. Brands must be able to demonstrate exactly how their AI reached its safety conclusions, especially as claims substantiation is becoming more rigorous globally.
AI can help by matching claims to appropriate evidence types, like clinical studies for objective claims and consumer testing for subjective claims, but the underlying science must be valid.
CDU: As AI systems rely more heavily on consumer data, biometric inputs, and behavioral signals, what governance or ethical challenges should beauty companies be preparing for now?
Wayne Liu: The governance challenge of 2026 is about native design trust and holistic transparency. We are moving into an era of ‘Privacy-by-Design,’ where data protection is embedded into the architecture, design, and deployment, not just its legal terms.
First, we must champion granularity in consent architecture. I believe biometric analysis must be strictly opt-in and purpose-specific. Consumers shouldn’t just click the ‘agree’ button to a policy; they should affirmatively choose to engage with a specific feature.
The ‘consumer-first’ approach should move from optional to a business imperative.
Second, we face the challenge of Algorithmic Fairness. As AI moves into upstream formulation, companies must ensure their models are trained on inclusive datasets. If an AI is making formulation decisions, it must represent all 90,000+ skin tones we track to avoid ‘digital bias’ in product efficacy.
Finally, beauty companies must prepare for a de-fragmented global standard. Rather than a patchwork approach, the leaders of 2026 will adopt a ‘ceiling, not a floor’ strategy, applying the strictest global standards, like the EU AI Act, across all jurisdictions.
The goal is radical transparency: if a consumer asks why a recommendation was made, we should be able to provide the ‘explainability’ behind the algorithm.
CDU: If you were advising a mid-size beauty manufacturer today, what would be the single most strategic AI investment, or capability, they should prioritize?
Anastasia Georgievskaya: The most strategic investment is AI that improves decision-making around product performance and consumer outcomes, not AI that simply automates engagement. This means investing in systems that combine baseline skin measurement, validated scientific data, and real-world feedback to understand how products actually work.
From our perspective, tools that support formulation validation, claims substantiation, and accurate product matching deliver long-term value. They help brands reduce waste, improve consumer satisfaction, and build trust.
As consumers increasingly shop for results, evidence-based AI will matter far more than surface-level personalization or short-term conversion tools.
CDU: Looking ahead to 2026, how might AI reshape collaboration between brands, ingredient suppliers, and contract manufacturers across the value chain?
Jane Yoo, M.D.: AI will push the industry toward shared standards and faster iteration, but only if all partners agree on standardized data practices.
- Brands will use AI to generate tighter specifications; CMs will use AI to validate manufacturability earlier.
- Suppliers will package not just marketing claims, but structured data on stability, irritation risk, and compatible systems.
- Digital twins for scale-up: More simulation of how a lab formula behaves in production—reducing failed scale-ups.
- Increasing demand for provenance (sourcing, contaminants, allergens, impurities) with data systems that can be audited.
This will lead to higher requirements for suppliers to provide clean, standardized data.
