Image: The first image of a black hole, taken in 2019. Image source: Event Horizon Telescope
Data can be collected faster and processed more accurately. Unmanned space missions can operate reliably without constant human supervision. If ethical and safety concerns are adequately addressed, the accelerating development of artificial intelligence could be a boon for space science.
In recent decades, artificial intelligence has brought about an unprecedented revolution. The speed at which artificial intelligence has developed in science, business and other social sectors is viewed with increasing astonishment and interest by some, and with skepticism by others. Who among us does not feel disturbed when seeing a 3D image of the deceased Sheikh addressing the climate summit in Dubai? In early 2023, panicked reactions from the science and technology sectors led to an open letter calling for a pause (Pausing Giant AI Experiments: An Open Letter2023).
We cannot imagine what could happen when the development of artificial intelligence begins to outpace that of humans, or – which is even more terrible – when the machine completely takes over responsibility for humans? But the rapid development of artificial intelligence also has advantages. Many scientific fields benefit greatly from the new possibilities offered by artificial intelligence. This is also the case in my specialty, space research, which has always been at the forefront of using advanced technologies and innovative methods. In space research, AI is not only helping to make existing processes more efficient, but also opening up entirely new possibilities that previously seemed out of reach.
Take, for example, the processing and analysis of vast amounts of data collected by satellites, probes, and telescopes. Traditionally, processing this data would take an enormous amount of time. However, with artificial intelligence, machines can identify patterns that were previously unnoticed. In fact, they can make independent decisions based on the information collected. Data is processed faster and more accurately.
The European Space Agency (ESA) is currently testing new artificial intelligence concepts and computer algorithms in new smart satellites. For example, the first AI-CubeSat, a type of small satellite launched in 2020, can itself identify relevant images to send to Earth, so that the bulk of the images do not have to travel across the galaxy. Thanks to machine learning algorithms, astronomical objects such as galaxies are being classified at a previously unimaginable rate. Researchers can now navigate through large data sets much faster and focus on interpreting the relevant information.
A second example of the application of artificial intelligence in space research is the image of the black hole in the Messier 87 (M87) galaxy, taken by the Event Horizon Telescope (EHT) in April 2019. The EHT is a network of seven telescopes, but together they are unable to observe the hole. Black is completely in the heart of the galaxy. The pale orange donut seen in the first photo, the original image was AI whipped up to the thin gold ring in the second photo. The machine learning algorithm predicted that the missing data points would produce an improved image, with the half ring being narrower. This clarification has important implications for the theoretical models that will be launched on the black hole.
In addition to data collection and processing, artificial intelligence in space research is also playing an accelerating role in the development of autonomous space missions. Until now, the astronauts themselves and Mission Control have been the ones making decisions during spaceflight, but thanks to advances in artificial intelligence, spacecraft can increasingly operate autonomously. For example, the autopilot for SpaceX’s Falcon 9 rocket uses artificial intelligence when the rocket docks with the International Space Station (ISS). In addition, the launch paths and orbit are created using artificial intelligence, and the artificial intelligence helps monitor atmospheric environmental factors and optimize fuel use. Suddenly, expeditions to distant planets, moons, and other celestial bodies became possible without constant human supervision.
NASA’s Mars exploration program is actually an example of this. Due to the difference in distance between Mars and Earth, there is a delay in transmitting signals between the two planets. Therefore, robotic rovers on Mars, called Mars Exploration Rovers (MERs), must be able to perform their daily work semi-autonomously. Therefore, it is equipped with advanced artificial intelligence systems that help it avoid obstacles, operate scientific instruments, and choose the best route. The MER rover Curiosity has been orbiting Mars since 2012. As part of research into the red planet’s habitability, it searches for traces of life and studies the geological processes that shaped the landscape. Since 2020, the MER rover “Perseverance” has also been roaming the surface of Mars. It is used as an active laboratory and collects samples that will be sent to Earth. Thanks to artificial intelligence, these missions can achieve their scientific goals more efficiently and with greater flexibility.
However, there are not only advantages to incorporating AI into space research. One of the most pressing issues surrounding autonomous systems in space is the ethical one, as the authors of the above-mentioned open letter have already written. It is possible to give machines the power to make decisions, but who is responsible for the unforeseen but conceivable consequences? Many satellites collect tons of data about the Earth. In principle, they can “see” everything, including data that people or organizations would prefer not to publish or disclose. Machines typically pay little attention to privacy laws. Strong ethical guidelines are therefore needed that require autonomous systems to operate in accordance with human values and standards. It also works in reverse. Within the framework of “space law”, consideration is given to the rights of the universe itself. Ultimately, man cannot imagine himself simply as the “owner” of the moon.
Besides the ethical issue, there is also the issue of safety. For example, it is necessary for the spacecraft to enter the atmosphere at a very precise angle. The slightest deflection could cause instability or overheating of the capsule, with serious consequences for the astronauts on board. But the ability of artificial intelligence to make independent decisions depends on the quality of the algorithms used. To prevent accidents as much as possible, it is necessary to carry out comprehensive testing and adequate safety measures. But if we approach this ever-evolving technology responsibly and preparedly, we can reveal the secrets of the universe and their consequences for humans in ways previously unimaginable.
This article also appeared on MaJa, het Youth Academy Magazine.
References
- Medeiros, L., Psaltis, D., Lauer TR, Özel, F. (2023). Image of the M87 black hole reconstructed using Primo. Astrophysical Journal Letters, 947(1), L7.
- Pausing Giant AI Experiments: An Open Letter (2023, March 22). Future Life Institute. https://futureoflife.org/open-letter/pause-giant-ai-experiments/
- Smarter satellites: ESA discovery accelerates artificial intelligence in space (2022, May 12). European Space Agency. https://www.esa.int/Enabling_Support/Preparing_for_the_Future/Discovery_and_Preparation/Smarter_satellites_ESA_Discovery_accelerates_AI_in_space
- Event Horizon Telescope Collaboration. (2019). Results of the first M87 Event Horizon Telescope. I. The shadow of a supermassive black hole. Astrophysical Journal Letters, 875(1), for 1.
