AI for cosmetics: use it in the right places.
The industry sells through emotions and trust in results, so AI should be used purposefully, not just for convenience.
AI is suitable for creating moodboards, testing concepts, and campaign color tones internally before actual production. What you absolutely should not do with AI are before-and-after images, product effectiveness visuals, and any claims of benefits, as these pose both legal risks and undermine buyer trust. The simple boundary is: AI serves internal creativity, while public-facing results must be real.
- creating a mood board for the visual world to test the color tone of the campaign concept
- explore multiple creative directions internally
- the brand wants to quickly expand its aesthetic range
- before-and-after images and product result visuals
- Testimonials and model images implying real results.
- a premium brand positioning itself as unique to avoid looking mass-produced
The cosmetics and beauty industry sells through two things: emotions and trust in results. AI can support emotions in the internal creative process, but no tool can replace the truth when it comes to trust in results. This is why you need to understand what AI can assist with and what it should not touch before implementation.
The hardest boundary of the industry
With cosmetics, advertising must be truthful, and all claims about efficacy must be based on real evidence before being published. This means that before-and-after photos, images of product effectiveness, and testimonials should not create an impression of results that have never occurred.
Using AI to enhance before-and-after photos, creating perfect skin that doesn't exist, or simulating product effects that the actual product cannot achieve is misleading advertising. This is not just an aesthetic issue but a legal risk with real consequences.
Imagery implying a result or experience that has never occurred is misleading advertising. Before-and-after photos and product effectiveness must reflect reality.
FTC, warning on misleading AI advertising (2024)
Where is AI strong?
In the beauty industry, AI performs best in stages that are not public-facing and do not imply product results.
- Create a moodboard (visual inspiration board) and test the campaign color palette before actual production, helping to save time and approval costs.
- Visual language testing for a new product line: materials, lighting, spatial feel.
- Creating packaging sketches to present internal ideas before moving to actual sample production.
- Support in writing product descriptions and educational content about ingredients and skincare processes, as long as the content is truthful to the actual product.
When to choose which direction
Common errors
- Using AI to edit photos to the point where the results look far from the actual product is the most common mistake and carries the highest risk.
- To let AI lead the entire brand identity without anchoring to its own system. The result is often aesthetically pleasing but resembles many other beauty brands, as AI pulls from the most popular samples in its data.
- Overusing perfect faces and skin created by AI. Research shows that buyers are increasingly recognizing AI traces and feel less connected to those images, having a counterproductive effect in the emotional selling industry.
- Unclear about whether to use synthetic models (AI-generated models), leading to a lack of transparency with buyers and risks of unrealistic expectations.
Sinh Vũ's viewpoint
Sinh Vũ views AI in the beauty industry as a tool to expand creative boundaries during the exploratory phase, not a place to rely on for promised results. The approach is to anchor AI to an established brand system, allowing AI to serve your unique identity rather than pulling the brand towards the general aesthetics of machines.
The role of Sinh Vũ in these projects is to serve as the creative director maintaining visual standards and authenticity. The legal responsibility for product efficacy lies with you and must be validated with real data, not AI-generated images.
A beauty brand stands strong not because it looks the best on social media, but because buyers believe what the brand says is true.
Decision checklist
Topic: AI for cosmetics and beauty: using it in the right places, avoiding risks. Sinh Vũ guide, sinhvu.com
Select each item you find appropriate, then print or save as PDF to take with you.
If you have marked most of the signs above, this is the time to discuss in more detail. Sinh Vũ can help you review and propose a direction.
References
FTC Health Products Compliance Guidance; FTC warning on misleading AI advertising (2024); Generative AI Design and Credibility, Salted Stone (2023); practical experience of Sinh Vũ.
Frequently asked questions
Is it permissible to use AI-generated models?
Yes, but you need to carefully consider two things. First is image rights and consent, because even without real people, AI model images still imply beauty and can create unrealistic expectations. Second is transparency with buyers, as many discerning individuals can recognize AI traces and feel deceived. For a premium positioned brand, Sinh Vũ advises limiting the use of AI models as primary imagery.
Can AI help create a brand identity for cosmetics?
You can use AI to explore aesthetic directions, test color tones, and create initial mood boards. However, if you let AI guide the entire final result, there is a high risk that the identity will look like many other beauty brands because AI tends to pull from the most popular aesthetic samples. The role of the brand creator is to anchor AI to the system and the unique identity of each brand.