AI news ·
CAS launches Newton, an agentic AI tool built on 150 years of curated scientific data
CAS Newton, a new AI agent from the American Chemical Society, draws on 150 years of curated chemistry data to answer research questions. Three in four early users rated its answers more trustworthy than other AI tools.
CAS Launches AI Tool Built on 150 Years of Chemistry Data
The American Chemical Society's CAS division has released CAS Newton, an AI agent designed to answer scientific questions by drawing on more than 150 years of curated chemistry knowledge. The tool aims to give researchers faster access to verified information without sacrificing accuracy.
Early testing shows three out of four users find CAS Newton's answers more trustworthy than those from other AI tools. The difference lies in its foundation: instead of relying solely on general training data, CAS Newton grounds its responses in the CAS Content Collection, a database of rigorously reviewed scientific literature and data.
How It Works for Research Teams
CAS Newton operates as a conversational agent. It asks clarifying questions, retains context across multiple exchanges, and synthesises results across different research steps. This approach is meant to help researchers move from initial questions to answers they can verify and cite.
The tool is available within CAS SciFinder and CAS BioFinder, as well as through a standalone interface. R&D organisations can also integrate it into secure environments alongside proprietary data through APIs and third-party AI platforms, allowing teams to combine CAS's public knowledge with their own internal research.
Privacy and Security
CAS says user queries and results are not used to train models across other users. Input remains confidential, and the company applies what it calls an ethical AI approach to governance.
Tim Wahlberg, Chief Product Officer and interim President of CAS, said the launch "marks a major shift in how scientific knowledge flows through discovery," allowing researchers to move from questions to grounded answers without compromising trust.
Learn More
For researchers working with AI tools and agents, exploring Generative AI and LLM Courses can help build foundational knowledge about how these systems work. Those focused on research applications may also benefit from AI Research Courses.