Most cosmetics tested on animals will soon be illegal in California, under a new law signed Friday by Gov. Jerry Brown.

The ban, authored by Sen. Cathleen Galgiani (D-Stockton), will go into effect on Jan. 1, 2020, and will apply only to cosmetics tested on animals after that date. It allows exceptions when a federal authority, such as the Food and Drug Administration, requires animal testing.

The measure passed easily in the Legislature last month after late amendments reduced its scope, removing most opposition. Prior to those amendments, some opponents had worried the ban would hurt business in China, which requires animal testing for health reasons. The amended law allows manufacturers to test products on animals to meet foreign regulations, as long as those tests are not used to confirm a product’s safety in California.

The August amendments also specified that a manufacturer can sell cosmetics containing ingredients tested on animals, provided the animal testing was not done by the manufacturer or its suppliers. That provision addressed some companies’ concerns that other organizations could test their key ingredients on animals, stopping sales.

The law ultimately received widespread support from cosmetics manufacturers, celebrity advocates and animal rights groups.

As good as it is for preventing animals from being mistreated, I'm a little worried about what impact this may have on the health of people who use those products after that date.

While one ingredient may be perfectly fine, what it does to the skin when combined with another ingredient may be different. What are the possibilities that this can lead to some combinations that cause harm to animal (and thus possibly human) skin being released because they were unable to test its effects?

Unless there's a step in the process I'm missing, in which case please do let me know.

Are you asking of whether there is or isn't another alternative? Because they are plenty alternatives that animal-cruelty free brands use with still as much success as other companies that do use animal-testing to test their products. It's hard to balance idealism and pragmatism in regards to so, but there are still effective alternatives that don't have to results to testing on animals (i.e computer models / simulations, tissue models and cell based / genetic testing methods) to a higher standard as they're bound to get a more suited result, especially being the mad difference between human & animal skin, and chemical reactions.

What do we know about the computer models / simulations? The last I remember involving computer models and simulations was when protein strand building was publicly sourced through a game (the Foldit game). However, that was done with lots and lots of people playing that game, and not a computer calculating it for them. That was quite a few years ago (I think about 7 years?), so I'm not sure that we've quite gotten computers up to spec for something like that yet. I also wouldn't trust the data from that computer if that was the only data being provided (as opposed to having other sources of data, like tests on tissue models like you were referring to).

Within the initiative “FP7-HEALTH-2010-Alternative-Testing-Strategies” the project NOTOX will develop and establish a spectrum of systems biological tools for toxicity prediction. These include experimental and computational methods for i) organotypic human cell cultures suitable for long term toxicity testing and ii) the identification and analysis of pathways of toxicological relevance.

Systems biological tools in NOTOX comprise experimental methods using organotypic human cell cultures and computational tools. NOTOX will generate own experimental data to develop and validate predictive mathematical and bioinformatic models characterizing long term toxicity responses. Cellular activities will be monitored continuously by comprehensive analysis of released metabolites, peptides and proteins and by the estimation of metabolic fluxes using 13C labelling techniques (fluxomics). At selected time points a part of the cells will be removed for in-depth structural (3D-optical and electron microscopy tomography), transcriptomic, epigenomic, metabolomic, proteomic and fluxomic characterizations. When applicable, cells derived from human stem cells (hESC or iPS) and available human organ simulating systems or even a multi-organ platform developed within the SEURAT cluster will be investigated using developed methods.

Together with curated literature and genomic data these toxicological data will be organised in a toxicological database. Physiological data including metabolism of test compounds will be incorporated into large-scale computer models that are based on material balancing and kinetics. Various “-omics” data and 3D structural information from organotypic cultures will be integrated using correlative bioinformatic tools. These data also serve as a basis for large scale mathematical models. The overall objectives are to identify cellular and molecular signatures allowing prediction of long term toxicity, to design experimental systems for the identification of predictive endpoints and to integrate these into causal computer models.

I actually do think that the tissue model part might be useful, but it sounds like it's expensive, especially in comparison to the purchase of, say, a hundred adult pigs. While it sucks for the 100 pigs (especially if it turns out that the cosmetic is harmful and that they got it wrong), I feel like it's better than the millions of customers who would be wearing those dangerous cosmetics when the companies decide to cut corners because of testing costs. I feel that limiting the means of testing these companies have (especially before we have readily available and open sourced technology to allow them the other alternatives) is just going to make them put those customers at risk, especially if their only alternative is to not sell in California, and thus lose a large amount of potential buyers. . . part of the cells will be removed for in-depth structural. . . transcriptomic, epigenomic, metabolomic, proteomic and fluxomic characterizations. . .