In this month’s column we take a look at the challenges faced with the bullet train of AI entering society, business, politics and of course cosmetics…
The shenanigans at the top of OpenAI have been widely reported in the media over the past week or so. Who knows, what that whole thing was about, and it might even end up becoming this year’s Christmas movie? One thing is certain: artificial intelligence (AI) is here to stay, and moreover, humans must continue to be its primary controllers. The use of avatars in advertising (though not always explicitly stated), apps and other tools to determine whether a product is “toxic” (via an endorsement label), and attempts by these entities to assert their superiority over the authority and even the dignity of regulatory bodies, qualified toxicologists, safety assessors, and SCCS toxicologists, are just a few examples of how artificial intelligence has infiltrated our industry.
Labels with seals that promote and endorse goods obtained through these Apps, and other ‘unapproved’ online certifiers, are becoming more and more common. However, even while some seals and endorsements may have genuinely charitable intentions, their primary goal is to boost sales; they are businesses after all. It is always advisable for brands to conduct consumer insight studies to determine the true value of these seal/endorsement Apps and companies, as well as to assess the credibility of the seal, call them out when necessary, and balance the investment value of these labels against the return on investment for the final product or brand.
In terms of AI, these seals and labels are linked to data obtained from the internet, which contains a great deal of false information regarding the substances in cosmetics, sentimental social networking platforms, and a highly perplexed and irate customer base, including ourselves. Regrettably, a lot of misinformation is mostly produced by the industry. There is a huge risk when marketers and copywriters use artificial intelligence (AI) to write their claims and are confronted by their regulatory affairs team or even legal teams. Frequently, the bean counters come out on top. People with some grey cells can judge the vast quantity of false material on the internet as fact or fiction. Since AI systems are not human, they cannot yet distinguish fact from fiction; at the very least, they “stumble” when certain search terms are entered. This is mostly because they are algorithms, and the rate-limiting step is the “wording” produced by AI systems. Even ChatGPT has a disclaimer stating that fact-checking by a human is necessary because the information is not wholly reliable.
One example, when we consider the claim ‘clean beauty’ — you only have toGoogle the term or type for the definition, and pages appear in the listing talking about so-called ‘harmful ingredients’. We have failed in educating journalists (sensational headlines rule!), influencers (even more sensational headlines), the consumer, marketing, advertising…and the list goes on. The fact that the key mission and metier of the cosmetic industry is to do no harm, and ‘harmful ingredients’ are banned anyway, seems to fall by the wayside. As there is no regulatory definition of ‘clean beauty’, the misinformation will continue to grow. This impacts any AI development. Moreover, as we have no international harmonisation for cosmetic regulations — even the new FDA’s MOCRA has no Annexes on banned or specifically controlled ingredients — then we are in for a rough ride, at least for the foreseeable future.
On the other side we have professionals, e.g., the LinkedIn thread (with thousands of ‘reads’), highlighting a British newspaper publishing a TikTok trend by a US pharmacist advocating the mixing of two pharmaceutical products meant for vascular treatment and sinus treatment to create a finger-mix to rub around the eyes to get rid of eye bags. With this type of carelessness AI will struggle to keep up. Other examples include the ingredient suppliers themselves. The list of claims some active ingredient suppliers present are either pharmaceutical in nature, are overtly dermal in nature, not permissible under current EU or even other regulations, and because there is no ‘claims’ regulation enforced on the supplier, they continue to feed misinformation to the brands, who ultimately become confused, and spread the misinformation more widely. The science has to be translated accurately in a consumer-meaningful way, be relevant, and compliant.
Nonetheless, hope remains on the horizon, and without it, we wouldn’t be here. Like-minded colleagues and I, know what needs to be done; the challenge is how to initiate it? Gradually, we will become influencers over social media hysteria, journalists and their sensational headlines, non-compliant claims, etc. It has been said to me more than once by a member of the cosmetic industry that I can’t change it, and it will not change because we have to maintain the status quo. It is true I cannot change it alone – I do not intend to, one reason being is my health. Yet, those of us in the industry with common sense and a will to do right by the industry, can change it and the status quo — together. As a start, by chance this week, I was contacted by one of my Scandinavian customers with whom I have enjoyed working these last few years, introducing me to some exciting guys based in the USA, and their company Sei. They have developed an AI tool (ironically for the Finance sector) to aid Regulatory affairs managers, legal teams, Responsible Persons, and even regulatory bodies themselves. The tool judges your claims based not on internet misinformation and hysteria, but on genuine fact — regulations (e.g., EU, FDA, etc.), judgements from courts and authorities, etc. It even gives you the reason why your claims are not permissible and the requirement for ‘evidence’. Moreover, if the tool develops in the way it should, it will help put an end to misinformation, scaremongering, and me-too-toxicologists. You will hear more from them on this AI tool next year, and no doubt from me too!
Theresa Callaghan
Theresa Callaghan, a PhD biochemist with over 35 years of experience in corporate skin care research, has held key R&D senior roles for companies including LVMH, Unilever, Marks & Spencer, J&J, Evonik, Hill-Top Research, and proDERM. In 2008, she created Callaghan Consulting International, focusing on cosmetic claims development with brands and ingredient suppliers. She is a widely published author, frequent speaker, and contributor to peer-reviewed journals. Her acclaimed book, 'Help! I'm Covered in Adjectives: Cosmetic Claims & The Consumer', has gained popularity. She is a member of the Society of Cosmetic Scientists and British Herbal Medicine Association, and has lectured at the University of Sunderland's School of Pharmacy and Cosmetic Sciences. Theresa serves on the editorial peer review board of the International Journal of Cosmetic Science. She also mentors; advises TKS Science Publisher; and writes monthly for BEAUTYSTREAMS Ingredient Pulse; and has her own Cosmetic Claims Insights Column with EuroCosmetics.
