Artificial Intelligence: US FDA Outlines 7 Steps To Establishing Model Credibility

The risk-based framework described in a new draft guidance starts with defining the question of interest and context of use and includes development and execution of a credibility assessment plan. The guidance is limited to AI models used to support regulatory decisions about drug safety, effectiveness or quality.

Artificial intelligence
The FDA outlined a seven-step framework for establishing credibility of artificial intelligence models. (Shutterstock)
Key Takeaways
  • A new FDA draft guidance describes a seven-step, risk-based framework for assessing and demonstrating the credibility of artificial intelligence models when outputs are used to make drug regulatory decisions.
  • The guidance also discusses model life cycle management and FDA engagement opportunities.
  • The agency seeks feedback on whether the proposed assessment framework aligns with industry’s experience, and whether the available agency engagement options for sponsors and other interested parties are sufficient.

The US Food and Drug Administration wants drug sponsors to follow a seven-step framework for assessing and establishing the credibility of artificial intelligence models when their outputs are used to...

The framework employs a risk-based approach that generally will be tailored to the specific context of use and the model’s risk, the FDA said in a new draft

Read the full article – start your free trial today!

Join thousands of industry professionals who rely on Pink Sheet for daily insights

  • Start your 7-day free trial
  • Explore trusted news, analysis, and insights
  • Access comprehensive global coverage
  • Enjoy instant access – no credit card required

More from AI

More from Advanced Technologies

EU Health Data Space May Speed Up R&D Through Access To Multi-Omics & Clinical Record Data

 

The European Health Data Space framework will allow companies to accelerate R&D processes and identify new molecular targets faster by facilitating centralized access to certain types of high-quality data, Finland’s Orion Pharma says.

EU Wants Industry To Define ‘Trade Secrets’ Under Health Data Sharing Regulation

 

The European Health Data Space Regulation is deliberately “vague” when it comes to defining trade secrets because the EU wants the pharma industry to make recommendations on safeguarding intellectual property, a policy officer for the European Commission says.

Tougher Approval Standards May Follow Vinay Prasad’s Appointment To Lead US FDA’s CBER

 

Industry is concerned that Prasad may make regulatory flexibility tougher to obtain for cell and gene therapy, while vaccine and public health advocates are angry about Prasad’s criticisms of US COVID-19 policies.