Artificial Intelligence (AI) has emerged as a transformative force in the space sector.
AI technology has significantly improved satellite navigation and control systems, leading to heightened precision and efficiency. Earth Observation (EO) is also impacted by AI: by using AI algorithms, satellites can analyse extensive datasets concerning climate patterns and natural disasters, for instance. And, when it comes to satellite communications, AI can, among others, optimize communication protocols as well as facilitate network management. AI is also increasingly relevant for satellite operations: by employing advanced algorithms and machine learning methods, satellites can autonomously determine their positions, adjust orbits, and navigate through complex space environments. And space exploration is also benefitting from the use of AI technology.
Worldwide, the opportunities brought by AI, but also its risks, are being widely discussed. Countless initiatives, from policies, programmes and regulatory frameworks, are being put forward to ensure that the development of AI is not hindered whilst at the same time to ensure that its risks are mitigated.
In the European Union (EU), the European Commission (EC) proposed the AI Act in 2021, with a view to establishing a set of rules for placing on the market, putting into service and using AI systems in the Union, further laying down prohibitions on certain AI practices, among others. It followed a set of EU policy documents, notably, the Communication AI for Europe, the Communication Fostering a European Approach to AI, the Coordinated Plan on Artificial Intelligence, the White Paper on AI, and the AI-HLEG Ethics Guidelines for Trustworthy AI. Some of the above documents already acknowledged the role of AI in a set of industries, including in space, further recognising the role of data generated by the space sector for training AI algorithms. The space sector was also highlighted as one of the sectors where research and innovation on AI should be focused. The Communication Shaping Europe’s Digital Future also referred to the role of AI in an increasingly digitized society, whilst the 2030 Digital Compass underlined EU leadership on ethical AI. In this scope, and in its Path to the Digital Decade, a target of at least 75% of Union enterprises having taken up AI (among others) is set.
The AI Act is currently still under negotiations in the EU, following a new version put forward by the European Parliament in June 2023.
Given the cross-sector role of space products and services, the AI Act is expected to impact the space sector. For instance, high-risk systems are subject to a set of mandatory requirements aimed to ensure that they do not pose unacceptable risks to important Union public interests. In this regard, “AI systems intended to be used as safety components in the management and operation of road traffic and the supply of water, gas, heating and electricity” – where space services, such as GNSS, can be used – are considered high-risk systems.
In addition to AI-specific frameworks, others are relevant when it comes to AI, in light of the challenges brought by it, notably on privacy, transparency and explainability, fairness and non-discrimination, cultural sensitivity, safety and security, human control and accountability, sustainability (notably as relates with consumption and emissions, and ESG reporting) and AI liability, including when it comes to liability for defective products. Provisions on products (such as on machinery products aimed to ensure health and safety protection in light of new risks arising from emerging technologies, notably due to the evolving and autonomous behaviour of machinery), and on intellectual property, are also relevant. As are space specific legal frameworks, including existing applicable national laws, and the UN Space Treaties, including notably when it comes to responsibility and liability for space activities.
The synergies between space and AI are also acknowledged, for instance, in the EU Space Strategy for Security and Defence, which notes that space shall be “more systematically taken into account in relevant EU policies and initiatives, such as on quantum technologies, or artificial intelligence”. The EU Space Programme Regulation also mentions AI, noting its role in EO.
Likewise, entities such as the European Space Agency (ESA) are increasingly investing in AI: for instance, in 2022, ESA Discovery allocated funding to 12 projects aimed at investigating the potential utilization of cutting-edge AI advancements and advanced computing paradigms to enhance the responsiveness and autonomy of satellites. And ESA further launched, together with partners, AI4EO, aimed to promote the leveraging of AI technology in Earth Observation data. ESA is further investigating the use of AI in GNSS systems.
In turn, in October 2023, the European Union Agency for the Space Program (EUSPA) announced Cassini Challenges, a competition that tasks innovators with developing cutting-edge solutions to solve today’s societal issues using satellite data/services from Galileo and Copernicus. One of the three thematic challenges is the Emerging Technologies Challenge: innovators must develop disruptive solutions combining EU space data with deep technologies such as AI.
Several EU countries have also developed approaches to AI, such as Portugal (notably under AI Portugal 2030), Spain, France, Germany, Italy, Luxembourg, Estonia, among others. Many of these EU countries are also actively using AI when it comes to satellite products and services. Likewise, other countries are investing and assessing AI: for instance, the UK has approved its National AI Strategy in 2021; the US has approved the National Artificial Intelligence Research and Development Strategic Plan, the Enterprise AI Strategy, and an Executive Order on AI (with a proposal for an Algorithmic Accountability Act having been put forward); China has put forward its New Generation AI Development Plan and its Governance Principles for New Generation AI; Japan approved its AI Strategy in 2022; and India has also approved a National Strategy for AI. As an example, India used AI in the recent landing on the south pole of the moon to predict potential obstacles or hazards in the lunar terrain and thus enable a secure landing.
Likewise, in Africa, the African Union Development Agency is formulating the African Union Artificial Intelligence Continental Strategy for Africa, whilst the Digital Transformation Strategy for Africa (2020-2030) already addresses AI. African countries are also developing approaches to AI, such as Mauritius, Egypt, Kenya, Tunisia, Rwanda, Nigeria and South Africa.
Collaboration between the public and private sector for AI in space is also increasingly relevant. For instance, IBM has recently launched an open-source geospatial artificial intelligence foundational model specifically designed for Earth observation data resorting to NASA repository of EO data.
In Africa, the Rwanda Space Agency (RSA) and STAR.VISION Aerospace announced the successful culmination of the Algorithm Rideshare Program, under which the satellite WJ-1A can now autonomously analyse images, obviating the necessity for data transfers and streamlining the assessment of land usage in real-time. And, in Angola, the Angola’s National Space Programme Management Office (GGPEN) and ENDIAMA, an Angolan company dedicated to prospecting, exploring, and selling diamonds, launched an advanced Earth observation tool that integrates satellite imagery, artificial intelligence, and Chat GPT.
The increasing importance and use of AI in the space sector shall, in any case, be aware of the legal and regulatory applicable provisions, with a view to ensuring not only legal compliance but also risk mitigation, thus ensuring a safe, secure and sustainable space activity.