Seqera AI: Your privacy matters
Last updated: Dec 18, 2024
We will not use your specific questions, samples, pipelines, anything else that you may paste into our AI solutions for downstream model fine tuning
Table of contents:
At the Nextflow Summit Barcelona 2024, we launched Seqera AI - the bioinformatics agent purpose-built to support scientists throughout the R&D lifecycle, and we’ve already seen remarkable product usage. As interest in Seqera AI continues to grow, we understand that data privacy is a top priority, and we’d like to provide you with clarity and visibility into how we handle the data we collect to reassure you that your information remains safe with us.
Seqera is not in the business of building foundational models; instead, we use, fine-tune, and develop applications with frontier models from partners such as OpenAI and Anthropic. We use the models from the providers, paired with application logic, to deliver a superior experience for developers and bioinformaticians. Per our agreements with these providers: NO INFORMATION IS USED IN THE TRAINING OF SUBSEQUENT MODELS OR SERVICES. We are bound by commercial contracts that we have entered with these corporations to guarantee your data privacy. Should we engage additional partners to provide our services, we will update this page to include them and keep you informed.
More information on the commercial contracts with OpenAI and Anthropic can be found at:
- https://www.anthropic.com/legal/commercial-terms - see section A.4: Anthropic may not train models on Customer Content from paid Services.
- https://openai.com/policies/business-terms/ see section 3.2: We will not use Customer Content to develop or improve the Services.
What data do we collect? and from whom
We collect general sentiment information, prompts, and chats for a period of 400 days for the purpose of application optimization. This information helps guide decisions on how much synthetic data should be generated for certain topics from public datasets in order to improve the quality of our service. Again, NO CUSTOMER INPUTS ARE USED IN SUBSEQUENT FINE-TUNING OR DIRECT MODEL IMPROVEMENT.
We optimize towards general bioinformatics and Nextflow use cases. Here is an example of some of the areas that we consider:
- How well does Seqera AI respond to questions about conversion from DSL1 to DSL2
- How well does Seqera AI address samtools errors
- Which prefilled suggestions by Seqera AI are the most popular?
Our pledge
Seqera AI is a purpose-built bioinformatics service dedicated to enhancing the quality of bioinformatics code and improving the average bioinformatics experience. We understand that you trust us with significant intellectual property and do not want that IP shared in the public domain. We are committed to upholding that trust in our services and will do our best to provide transparency on the data we collect, the customer cohorts it pertains to, and the partners we work with to deliver our industry-leading AI.