Expert Systems for Predicting the Bioavailability of Sun Filters in Cosmetic Products, Software vs. Expert Formulator: The Benzophenone-3 Case

There are only a limited number of molecules in a cosmetic formulation, which can passively cross the stratum corneum and be absorbed into the skin layers. However, some actives should never cross the skin in large concentrations due to their potential for side effects, for example, sunscreens. Artificial intelligence is gaining an increasing role as a predictive tool, and in this regard, we selected the Formulating for Efficacy (R) Software to forecast the changes in bioavailability of selected topical cosmetic compounds. Using the Franz diffusion cell methodology, various oils were selected as those with low release capability, and these were compared to those suggested by the software in Benzophenone-3-containing formulations. The software was able to predict the lipophilic phases, which, if utilized in the emulsion, were stable and sometimes even more pleasant in appearance and consistency than the reference emulsions prepared by the formulator. To date, however, Formulating for Efficacy (R) Software still has limitations as far as predicting the hydrophilic phase, as well as not being able to choose the emulsifier or the preservative system.