With so many industries now utilizing Artificial Intelligence (AI), where does its place lie in the realm of cosmetic formulation? What are the pros and cons of AI for cosmetic formulators?
Equally important: what are the cons of AI that can possibly impact the future growth and innovation of cosmetic formulation, and where do the real opportunities for AI lie for formulators in the personal care sector?
This is such a big topic that it has been split into a 2-part instalment. This first instalment will focus on the general pros and cons of AI in key areas such as preliminary research into which ingredients get used in certain types of cosmetic formulas and how much of each ingredient should be used.
In the next instalment, we’ll look at the major cons of using AI, the misinformation that impacts AI results on cosmetic ingredient safety, as well as the big opportunities AI can present for cosmetic formulators and chemists.
Pro: AI can be used to research consumer attitudes toward ingredients.
AI can be a useful tool when a cosmetic formulator is first investigating why consumers are navigating toward certain ingredients and potentially avoiding others. It can help provide insight as to consumer trends and opinions about what certain cosmetic products ‘should’ contain, based on what it can find on the internet.
For example, I ran the question*: ‘what ingredients should a serum contain?’, and received a long list of ingredients that is obviously used in other personal care products.
It listed them in categories of hydration (hyaluronic acid, glycerin, betaine and panthenol); brightening and antioxidant protection (vitamin C, ferulic acid, vitamin E and niacinamide); anti-ageing (retinol, peptides, AHAs and BHAs) and other helpful ingredients: ceramides, plant oils, snail mucin and propolis.
The cons of this search result:
While this search may have seemed like a helpful start, and to some degree it does provide a cosmetic chemist with a starting list of ‘go-to’ ingredients a consumer might need to see in their serum to consider it competitive with other offerings, the cons of these search results are:
- We are not given any indication about the regulatory limits related to the use of these ingredients, nor how much should be used, incompatibilities when combining them, or pH and formulation/stability needs.
- We have no idea on more unique cosmetic ingredient choices to offer a point of difference when formulating something new for a brand. It would take individual research on cosmetic actives available from various suppliers, which AI cannot do, to ensure an individualized marketing story can be created to appeal to our target market.
- It gave no information on how to construct the base serum formula for efficacy, safety or a suitable shelf life. It only listed cosmetic actives with no concrete information on their use or restrictions.
- It gives no technical information about the actual ingredient other than its INCI name. There are some cosmetic ingredients with the same INCI name but very different actions in cosmetic formulas due to varying molecular weights (for example, sodium hyaluronate, commonly referred to as hyaluronic acid) or grades (for example, many natural gums).
The final ruling: while AI can help a cosmetic formulator quickly search more popular materials used in current cosmetic products, it does not provide detailed information about which may be the best or even necessary choices to suit a specific formulation brief and provide a marketable point of difference.
Pro: AI can be used to run preliminary checks on inputs of cosmetic ingredients.
While I’ve listed this as a pro, the information that AI can provide on cosmetic ingredient input needs to be considered very carefully. AI will use the information on the internet to form a response, which may or may not be correct based on how much misinformation is already out there. It can give you a starting point on how much of an ingredient should be used, but this still needs to be confirmed by formal cosmetic training and formulation knowledge, supplier technical information, and regulatory checks.
For example, I ran the search*: ‘how much phenoxyethanol should I use in a cosmetic formula?’ and obtained some useful information, as well as some incorrect results:
- 1% or less is considered safe by EU regulatory authorities, CIR (Cosmetic Ingredient Review) panel and SCCS opinions. (Author note: correct; although it doesn’t explain what is the ‘best’ input for specific use in a certain type of formula.)
- It is often used in conjunction with other preservatives such as ethylhexylglycerin to obtain broad spectrum preservation in cosmetic products. (Author note: correct, but it does not explain why other materials are needed, which are best to use nor how much should be used).
- It lists the effective pH range to be 4 – 10 but it performs best at pH 6. (Author note: not totally technically correct; it needs to be used in conjunction with another material to ensure broad spectrum coverage and that will impact what the best final pH will be.)
- AI says use above 1% may cause allergic reactions or skin irritations. (Author note: phenoxyethanol is not a known allergen but may cause skin irritations at higher inputs. However, the regulatory limit is not mentioned which is 1%w/w in most regions.)
Cons of this search result include:
- While 1% is mentioned, it is not qualified by how that 1% is measured. It would be more accurate to list 1%w/w, as we should always do, when talking about inputs of cosmetic ingredients for formulas.
- More information about what phenoxyethanol is used with would impact how much should be used and what would then be the best final pH for the formula.
- It implies phenoxyethanol is an allergen which is not correct.
The final ruling: while AI can help a cosmetic formulator run preliminary checks on the recommended input for specific ingredients, the information found should not be relied upon. Moreover, if checking for preservatives or active ingredients, accurate sources and formal training should be relied upon instead, to ensure appropriate safe inputs are used in the correct formulation conditions for required results.
Is AI useful for cosmetic formulators?
Right now, AI should not be relied upon by a cosmetic formulator for cosmetic ingredient inputs or cosmetic formulation guides. It can however be used as a good starting point for market research and to understand consumer perception about certain cosmetic ingredients.
AI is not at a stage where it can be utilized to make decisions, but instead, provide preliminary information to be considered when making early ingredient selections towards building cosmetic formulations.
Get ready for the next instalment in this exiting topic, when we look at the cons and opportunities that AI has for cosmetic formulators.
Happy formulating!
* Note: if you run similar AI checks of questions your results may vary depending on your search engine and location.
Feeling inspired?
Then why not visit one of the in-cosmetics events around the world?