Episode•May 29, 2026•35 min
AI at NASA
NASA Acting Chief Data and AI Officer Kevin Murphy discusses how artificial intelligence is supporting missions, research, and the future of human spaceflight. Episode 424.
Key Takeaways
- 1
AI at NASA is a tool for pattern recognition, not reasoning
AI is software that learns patterns from data rather than following hand-coded rules.
It is used to process massive datasets that would be impossible for human teams to analyze manually.
The agency emphasizes a 'human-in-the-loop' approach, ensuring experts validate AI outputs for high-consequence decisions.
- 2
Governance and safety are paramount
NASA uses a sliding scale for AI oversight: exploratory research has more freedom, while safety-critical or resource-impacting applications require rigorous validation.
The agency strictly prohibits using AI to generate imagery that could be mistaken for authentic mission photography.
Data security is critical; NASA prevents sensitive data (CUI, ITAR, EAR) from leaking into public AI models.
- 3
Data management is the foundation of AI success
Good AI requires high-quality data; the 'garbage in, garbage out' principle applies.
NASA is working to break down data silos across its 10 centers to create consistent patterns for AI training.
Treating data as a strategic asset is essential for achieving high-velocity decision-making.
Intro
- In this episode, NASA Acting Chief Data and AI Officer Kevin Murphy explores how artificial intelligence is being integrated into space exploration, research, and agency operations.
- Kevin Murphy serves as NASA's Acting Chief Data and AI Officer and the Chief Science Data Officer for the Science Mission Directorate, overseeing the agency's data strategy and AI implementation.
– Defining AI at NASA
Artificial intelligence is software that learns patterns from examples instead of following rules written by hand.
- Murphy explains that AI is not about machine reasoning, but rather sophisticated pattern matching at scale.
– Governance and Use Cases
- NASA manages over 600 AI use cases, categorizing them by risk level to ensure safety and security.
As you kind of get to a situation where you can have kind of rights impacting, safety impacting, or large other types of implications, we have more rigorous testing.
– Real-World Applications
- ExoMiner: Using AI to identify exoplanets from Kepler and TESS telescope data.
- Mars Rovers: Autonomous route planning on Perseverance, allowing for longer traverses.
- Heliophysics: Foundation models to understand radiation environments.
– Ethics and Public Trust
- NASA maintains a strict policy against using AI to generate imagery that could be confused with real mission footage.
We cannot generate imagery through AI that could be confused as authentic mission images, for example. We are very clear in that.
Resources
Topics
Artificial IntelligenceData ScienceSpace ExplorationNASAMachine LearningAutonomous SystemsEthics in AI