Tom Branna, Chief Content Officer01.30.24
Artificial intelligence promises to revolutionize most industries. Repetitive and routine tasks such as data analysis and basic assembly line work are good candidates for AI intervention. Chatbots are already handling online customer service queries. But experts say AI can do more, much more. As reported on Happi.com, University of Miami researchers are using AI to improve raw material selection and formulation development.
At last month’s annual meeting of the Society of Cosmetic Chemists (SCC), Procter & Gamble Research Fellow Kukizo Miyamoto, PhD, detailed how one of the world’s largest fast-moving consumer goods companies is using AI to improve imaging and ultimately, improve products.
Miyamoto reviewed the technological improvements to facial imaging during the past 25 years, beginning with P&G’s Beauty Imaging System (BIS) Full Face Imaging Analysis (IA), evolving to Visia and Visia CR, and on to Magic Scan Contactless AI Diagnosis. In 2009, P&G’s Greg Hillebrand used the Visia system to detail wrinkle formation and in 2010, P&G’s SK-II brand began a 10-year, longitudinal study to understand wrinkle development. That research enabled P&G to create the Magic Ring Compact, a unique five-dimensional imaging analysis system.
More recently, artificial intelligence breakthroughs have enabled P&G to move from IA to AI. According to Miyamoto, AI enables researchers to predict the consumer’s No. 1, 2 and 3 skin issues. But does all this technology move the needle in terms of cosmetic breakthroughs?
“People who received the full face skin diagnosis after 10 years, looks better; 40s are the new 30s. 50s are the new 40s,” insisted Miyamoto. “That’s what we predicted when we began our research. AI enabled us to prove it!”
Besides developing better performing formulas, researchers are identifying better performing ingredients. Gravel AI is a data platform that helps skincare manufacturers grow their business with market insights. Its Chemical Demand Intelligence Platform for the personal care market provides chemical manufacturers with “instant” intelligence on where market opportunities lie; whether that opportunity lies in a particular chemical, country or application. According to Co-Founder and COO To Hong Chan, Gravel AI analyzes and 2.9 million products a month to process information.
“AI and Large Language Models (LLMs) bring the human touch to data analysis,” insisted Chan. “We created a three-step process to analyze marketing claims.”
Chan called it a breakthrough in marketing intelligence. The platform enables researchers to develop the most effective marketing claims based on ingredient and application combinations. At the same time, users can uncover trending ingredients.
“We can rank ingredients based on the greatest growth potential,” said Chan. “LLM and AI let us include marketing claims as an input factor in finding trending ingredients.”
What’s next? Chan said the platform will enable consumers to search for products using marketing claims with multi-language support. To get the best outcomes, he recommended using LLM with other AI/data science methods.
Haidong Liu, PhD, Schrödinger, explained how complex computational models help researchers understand the science and drive innovations in cosmetics and personal care formulations.
“Molecular modeling speeds research and development and reduces costs,” he insisted.
Liu detailed how Reckitt used Schrödinger molecular modeling to improve sustainability profiles of its packaging. At the same time, L’Oréal used digital simulation to explore sustainable product ingredient. Experimenting with new formulations in silico allowed L’Oréal researchers to make confident decisions far more quickly than if they were testing numerous new potential formulations in a laboratory—a process that often takes years to generate, according to Liu.
Hang Ma, PhD of the University of Rhode Island explained how his lab combines proteomics with AI to understand ferroptosis and other cell functions.
“We look at natural ingredients and how they impact skin and develop novel new ingredients for cosmetics and nutraceuticals,” he said. “We are always problem solving and digest information. AI can help us collect data and analyze information.”
Proteomics is the study of the interactions, function, composition and structure of proteins. Ferroptosis is a type of iron-dependent regulated cell death caused by unrestricted lipid peroxidation and subsequent membrane damage. Ma said ferroptosis plays a significant role in initiating UVB-induced inflammation in the skin. But to run 5000 traditional protein expression tests takes 13.2 years, according to Ma. Hence, the need for AI.
Ma’s lab discovered cannabinoids are excellent ferroptosis inhibitors—far superior to vitamin C, he added.
Now, his research is focused on finding more CBD-derived ferroptosis inhibitors. One such class is cannflavins, which have anti-inflammatory, anti-cancer, antioxidative and neuroprotective properties.
“Cannflavins are unique flavonoids from cannabinoids,” he concluded. “They show lots of promise, but we need collaboration!”
AI, too, shows plenty of promise; but for now, it won't eliminate the need for innovative, problem-solving cosmetic chemists.
At last month’s annual meeting of the Society of Cosmetic Chemists (SCC), Procter & Gamble Research Fellow Kukizo Miyamoto, PhD, detailed how one of the world’s largest fast-moving consumer goods companies is using AI to improve imaging and ultimately, improve products.
From IA to AI
Miyamoto reviewed the technological improvements to facial imaging during the past 25 years, beginning with P&G’s Beauty Imaging System (BIS) Full Face Imaging Analysis (IA), evolving to Visia and Visia CR, and on to Magic Scan Contactless AI Diagnosis. In 2009, P&G’s Greg Hillebrand used the Visia system to detail wrinkle formation and in 2010, P&G’s SK-II brand began a 10-year, longitudinal study to understand wrinkle development. That research enabled P&G to create the Magic Ring Compact, a unique five-dimensional imaging analysis system.
More recently, artificial intelligence breakthroughs have enabled P&G to move from IA to AI. According to Miyamoto, AI enables researchers to predict the consumer’s No. 1, 2 and 3 skin issues. But does all this technology move the needle in terms of cosmetic breakthroughs?
“People who received the full face skin diagnosis after 10 years, looks better; 40s are the new 30s. 50s are the new 40s,” insisted Miyamoto. “That’s what we predicted when we began our research. AI enabled us to prove it!”
AI and Personal Care Ingredients
Besides developing better performing formulas, researchers are identifying better performing ingredients. Gravel AI is a data platform that helps skincare manufacturers grow their business with market insights. Its Chemical Demand Intelligence Platform for the personal care market provides chemical manufacturers with “instant” intelligence on where market opportunities lie; whether that opportunity lies in a particular chemical, country or application. According to Co-Founder and COO To Hong Chan, Gravel AI analyzes and 2.9 million products a month to process information.
“AI and Large Language Models (LLMs) bring the human touch to data analysis,” insisted Chan. “We created a three-step process to analyze marketing claims.”
Chan called it a breakthrough in marketing intelligence. The platform enables researchers to develop the most effective marketing claims based on ingredient and application combinations. At the same time, users can uncover trending ingredients.
“We can rank ingredients based on the greatest growth potential,” said Chan. “LLM and AI let us include marketing claims as an input factor in finding trending ingredients.”
What’s next? Chan said the platform will enable consumers to search for products using marketing claims with multi-language support. To get the best outcomes, he recommended using LLM with other AI/data science methods.
Molecular Modeling
Haidong Liu, PhD, Schrödinger, explained how complex computational models help researchers understand the science and drive innovations in cosmetics and personal care formulations.
“Molecular modeling speeds research and development and reduces costs,” he insisted.
Liu detailed how Reckitt used Schrödinger molecular modeling to improve sustainability profiles of its packaging. At the same time, L’Oréal used digital simulation to explore sustainable product ingredient. Experimenting with new formulations in silico allowed L’Oréal researchers to make confident decisions far more quickly than if they were testing numerous new potential formulations in a laboratory—a process that often takes years to generate, according to Liu.
Data and AI/Proteomics
Hang Ma, PhD of the University of Rhode Island explained how his lab combines proteomics with AI to understand ferroptosis and other cell functions.
“We look at natural ingredients and how they impact skin and develop novel new ingredients for cosmetics and nutraceuticals,” he said. “We are always problem solving and digest information. AI can help us collect data and analyze information.”
Proteomics is the study of the interactions, function, composition and structure of proteins. Ferroptosis is a type of iron-dependent regulated cell death caused by unrestricted lipid peroxidation and subsequent membrane damage. Ma said ferroptosis plays a significant role in initiating UVB-induced inflammation in the skin. But to run 5000 traditional protein expression tests takes 13.2 years, according to Ma. Hence, the need for AI.
Ma’s lab discovered cannabinoids are excellent ferroptosis inhibitors—far superior to vitamin C, he added.
Now, his research is focused on finding more CBD-derived ferroptosis inhibitors. One such class is cannflavins, which have anti-inflammatory, anti-cancer, antioxidative and neuroprotective properties.
“Cannflavins are unique flavonoids from cannabinoids,” he concluded. “They show lots of promise, but we need collaboration!”
AI, too, shows plenty of promise; but for now, it won't eliminate the need for innovative, problem-solving cosmetic chemists.