Benefits and challenges of AI in space

February 27, 2024

Sallye Clark, Head of Space Law practice and Member, Mintz

Artificial intelligence (AI) is providing numerous benefits to society in the context of space. Developments in manufacturing, remote sensing, spacecraft control, and coordination will ease constraints to boost productivity and growth in the satellite industry. AI can drive down costs, increase innovation, mitigate space debris concerns, and guarantee certainty when operators find network coordination increasingly complex.

Benefits of AI in space

The beauty of AI is that it can mimic human intelligence through machine learning, by analyzing data independently or through autonomous embedded systems. Machine learning teaches machines to learn for themselves by ‘training’ a relatively simple algorithm to become more complex. Huge amounts of data feed into an algorithm, which adjusts and improves itself over time. Swarms of devices share their information in a network and learn from each device’s experience through hive learning. AI will progress from self-driving cars to autonomous navigation of spacecraft around Earth and other planets.

Examples of AI uses in space

AI is being used across four key areas: manufacturing, imaging, telemetry, and spectrum usage to perform tedious, time- consuming yet necessary tasks. AI is being used to identify areas of surveillance by learning to process and act upon signals it receives, disregarding substantial amounts of unnecessary data. AI is used to sort through Earth imaging data to monitor projects in remote areas, such as detecting pipeline leaks to measuring the impact of climate change:

  • AI is being used to monitor and control satellites. For example, SpaceX has implemented AI operations to avoid satellite collisions. AI could be used for other tasks, such as executing debris avoidance manoeuvres automatically. SpaceX is also using AI during take-off and landing of spacecraft to automate engine operations and manage functions such as deploying landing gear. This helps to optimise the use of fuel. Space X also uses an AI autopilot system to enable its Falcon 9 craft to carry out autonomous operations, such as docking with the ISS.
  • Lockheed Martin now has an “Operations Center of the Future,” which has the capacity to handle multiple space missions at once through automation, AI, and machine learning, to manage the rapidly increasing number (and complexity) of satellite constellations being deployed in an already crowded low Earth orbit.
  • NASA has set up an Artificial Intelligence Group that performs basic research that supports scientific analysis, spacecraft operations, mission analysis, deep space network operations and space transportation systems. NASA has also cooperated with Google to train its extensive AI algorithms to sift through the volumes of data from the Kepler exoplanet mission1. This led to the discovery of two new exoplanets previously missed by human scientists. AI is also being used on data from NASA’s TESS mission to identify candidate exoplanets.
  • AI is being used to simplify coordination for space-to- Earth transmissions to manage spectrum usage as the growing number of NGSO constellation becomes increasingly challenging. Satellites can learn to transmit using the appropriate frequencies and level of power output to avoid interference by detecting and avoiding co-channel interference at different stages of the satellite orbit. Adopting deep learning technology and automatic detection of transmitted frequencies from networks in proximity will reduce the interference burden for satellite networks.
  • The European Space Agency (ESA) is using AI in its space missions to enable rovers to autonomously navigate around obstacles, while data download from Mars rovers is scheduled using AI.
  • The German Aerospace Center (DLR) has been developing AI methods for space and earth applications for a number of years and in 2021 announced its intention to establish an institute for AI Security. Since 2018 it has supported astronauts onboard the ISS with an AI-enabled, voice- controlled interactive companion called CIMON (Crew Interactive MObile companioN).
  • French space agency CNES has worked with French company Clemessy2 to optimise the filling of rocket tanks using AI neural networks.
  • The UK Space Agency has funded a project involving the University of Southampton that uses AI3 to detect buried archaeological remains in satellite imagery providing construction companies with higher accuracy at an earlier stage. This will save them time and money during the planning permission process and help them reduce their carbon footprint.

Challenges of AI in space

Despite these many advances, the implementation of AI in space is not without its challenges. The widespread use of AI increases the risk of unauthorized system hacking, such as signal blocking, satellite takeover and destruction. To overcome these threats AI based systems will need to implement cybersecurity applications to protect against hacking.

While AI systems have clearly proven to be well suited for repeated tasks in harsh and hazardous environments, there is a potential mismatch between the AI’s current level of ability in perception and the intelligent decision-making that is required of humans when undertaking complex tasks in space. For AI to realise its full potential in space, we need to make advances in both autonomy and automation, thereby freeing humans to focus on tasks for which they are better suited.

We also need advances in robotic sensing and perception, mobility and manipulation, rendezvous and docking, onboard and ground-based autonomous capabilities. We will need human-robot integration and a suite of other data analysis tools for the foreseeable future to expand humans’ exploration of space.

The balance struck between human control and AI autonomy will depend both on the degree of social acceptance of the risks inherent in automation, and the level of cultural and political agreement about the human values and principles to be preserved in the development of increasingly autonomous systems.

Space law challenges of AI

Each legal jurisdiction will develop its own response to questions of liability, accountability, and responsibility where AI is applied to space. This will depend on the existing legal framework, including the specific features of criminal law, international humanitarian law, tort law, administrative law, etc. Current principles and rules in these fields may fall short in dealing with the interactivity, opacity, and unpredictability of autonomous “space objects.”

AI challenges of space law include those covered under the Liability Convention, a treaty that holds signatories to space related liabilities such as the damages to be covered, the procedure to be followed once such damages occur, or whether and to what extent the decisions of these “objects” fall under the fault of persons for whom a state is liable. These liability issues also concern the use of AI systems for space-based services, such as AI systems using Global Navigation Satellite System (GNSS) signals to support emergency response services, autonomous vehicles and unmanned aircraft systems.

  1. data-used-to-discover-eighth-planet-circling-distant-star/ ↩︎
  2. fluidic-systems-simulator-using-artificial-neural-networks/
  3. space-tech-to-help-solve-problems-on-earth ↩︎

Sallye Clark is a space and satellite attorney at Mintz who is highly regarded for her work on complex international satellite projects. She counsels her clients on legal, trade, and market access issues, represents her clients before regulatory bodies around the globe, and develops regulatory and legislative strategies for clients. You can reach her at [email protected].