The Viking missions showcased multiple spaceflight technologies that represented state-of-the-art capabilities: From digital line-scan imaging to the operation of complex onboard laboratories and software-controlled process autonomy. Since Viking, there have been extraordinary, and still accelerating, advancements in computing technology that impact science, society, and exploration. These developments have occurred in both hardware and software and have resulted in increasingly capable devices, advanced programming tools, and algorithmic innovations. The subset of artificial intelligence known as machine learning has emerged as one of the most transformative of these developments; it has major implications for space exploration and for improvements to the search for evidence of life beyond Earth. Those improvements include the integration of data across different scales and increased sensitivity to complex features in data, as well as the generation of adaptive strategies for sampling environments. In this article, the present and future nature of space exploration and astrobiological research is examined through the contextual lens of Viking and through the history and possible future of artificial intelligence.

The concept of a system-wide measure of the sustainment of life (habitability) for space-faring interplanetary species is introduced and explored. Although largely agnostic to the details of how interplanetary life might operate (e.g., via technology or by utilizing organism traits that are, as of now, unknown to us), some assumptions must be made about energy harvesting, orbital mobility costs, radiation risks, and resource requirements. A multi-modal figure of merit is developed for evaluating an interplanetary habitable zone (IHZ). An agent-based model is also developed to simulate the dispersal of interplanetary life in a planetary system and characterize the IHZ. For the solar system, resource weightings between planetary bodies dictate many overall behaviors, including the sequence of migration from Earth to the Moon, Mars, and asteroid belt. Comparisons with the Trappist-1 exoplanetary system also point to critical sensitivities in the balance between resource availability and risk or cost factors (e.g., radiation risks and orbital Delta-v costs) that determine the structure of an IHZ. Results suggest that our solar system may have an inherent, and significant, advantage for a space-faring species over a system like Trappist-1. This modeling approach may also have application to emerging space economies in our own solar system.

Advances in machine learning (ML) over the past decade have resulted in a proliferation of algorithmic applications for encoding, characterizing, and acting on complex data that may contain numerous multidimensional features. Recently, the emergence of deep-learning models trained across large datasets has created a new paradigm for ML in the form of Foundation Models (FMs). FMs are programs trained on large and broad datasets with an extensive number of parameters. Once built, these extremely powerful, flexible models can be utilized in less resource-intensive ways to build a variety of different downstream applications that can integrate previously disparate, multimodal data. The development of these applications can be done rapidly and with a much lower demand for ML expertise. Additionally, the necessary infrastructure and models themselves are already established within agencies such as NASA and ESA. At NASA, this work extends across several divisions of the Science Mission Directorate. Examples include the NASA Goddard and INDUS Large Language Models and the Prithvi Geospatial Foundation Model. Furthermore, ESA initiatives to bring FMs to Earth observations have led to the development of TerraMind. In February 2025, a workshop was held by NASA Ames Research Center and the SETI Institute to explore the potential of FMs in astrobiological research and identify the steps necessary to build and utilize such a model or models. Here, we share the findings and recommendations of that workshop and describe clear near-term and future opportunities in the development of a FM (or Models) for astrobiology applications. These applications would include a biosignature or life characterization task, a mission development and operations task, and a natural language task for integrating and supporting astrobiology research needs.

The application of convolutional autoencoder deep learning to imaging data for planetary science and astrobiological use is briefly reviewed and explored with a focus on the need to understand algorithmic rationale, process, and results when machine learning is utilized. Successful autoencoders train to build a model that captures the features of data in a dimensionally reduced form (the latent representation) that can then be used to recreate the original input. One application is the reconstruction of incomplete or noisy data. Here a baseline, lightweight convolutional autoencoder is used to examine the utility for planetary image reconstruction or inpainting in situations where there is destructive random noise (i.e., either luminance noise with zero returned data in some image pixels, or color noise with random additive levels across pixel channels). It is shown that, in certain use cases, multi-color image reconstruction can be usefully applied even with extensive random destructive noise with 90% areal coverage and higher. This capability is discussed in the context of intentional masking to reduce data bandwidth, or situations with low-illumination levels and other factors that obscure image data (e.g., sensor degradation or atmospheric conditions). It is further suggested that for some scientific use cases the model latent space and representations have more utility than large raw imaging datasets.

The potential discovery of life beyond Earth presents unique communication challenges for astrobiology. These include ambiguous data, public misconceptions, and the dynamics of social media platforms. Building on National Aeronautics and Space Administration's 2021 Standards of Evidence (SoE) workshop, a diverse group of experts-scientists, science journalists, content creators, and scholars-were convened during February and March of 2024 for the Communicating Discoveries in the Search for Life in the Universe workshop. This report summarizes structured discussions focused on how to responsibly share findings with different public audiences. Key themes that emerged from the workshop included the following: communicating uncertainty, reaching consensus, and building trust between the scientific community and the public. Such efforts will involve navigating the rapidly evolving landscapes of social media and academic (peer-reviewed) journal publishing. Workshop participants emphasized the need for proactive communication, early-career training in science communication, and interdisciplinary partnerships, all of which can foster sound public understandings of astrobiology research and its myriad of practices, mitigate misinformation, and sustain ongoing support for the search for life. In brief, this report includes the workshop rationale and structure, insights gleaned from past case studies and hypothetical future scenarios, common themes that emerged from the breakout groups, a discussion of the relationship of workshop outcomes to SoE, and guidance for individuals, agencies, and institutions. Key Words: Astrobiology-Science communication-Biosignature detection. Astrobiology 25, 743-758.

Computation, if treated as a set of physical processes that act on information represented by states of matter, encompasses biological systems, digital systems, and other constructs and may be a fundamental measure of living systems. The opportunity for biological computation, represented in the propagation and selection-driven evolution of information-carrying organic molecular structures, has been partially characterized in terms of planetary habitable zones (HZs) based on primary conditions such as temperature and the presence of liquid water. A generalization of this concept to computational zones (CZs) is proposed, with constraints set by three principal characteristics: capacity (including computation rates), energy, and instantiation (or substrate, including spatial extent). CZs naturally combine traditional habitability factors, including those associated with biological function that incorporate the chemical milieu, constraints on nutrients and free energy, as well as element availability. Two example applications are presented by examining the fundamental thermodynamic work efficiency and Landauer limit of photon-driven biological computation on planetary surfaces and of generalized computation in stellar energy capture structures (a.k.a. Dyson structures). It is suggested that CZs that involve nested structures or substellar objects could manifest unique observational signatures as cool far-infrared emitters. While these latter scenarios are entirely hypothetical, they offer a useful, complementary introduction to the potential universality of CZs.

The search for life elsewhere involves variables across multiple scales in time and space, often nested hierarchically. We suggest that the emergence of artificial intelligence learning systems offers critically important ways to make progress.

Studying the origin of life and its prevalence in the universe offers a perspective that compels us to look after our irreplaceable home in the cosmos. An entrenched conception of humans as distinct from Nature prevents us from seeing and embracing our place in space and time, to our catastrophic detriment. We call on our colleagues to harness this unique perspective to connect their research with broader problems facing humanity, while leveraging the trust, credibility, and privilege of the scientific enterprise.

I recently hiked a snow-covered trail renowned for its lack of cell service. Yet somehow, as I passed from one bend to another, a radio signal leaked into my almost-dead smartphone. Torn out of my reverie in the frigid air and under blue skies, without thinking I began scrolling through my messages. I’d received an urgent work entreaty, so I trudged back to my car and fired up my computer-controlled, hydrocarbon-combusting engine, and then I plugged in my 10-billion-transistor device and let it vigorously shuttle electrons. Only afterward, back on the trail, did I question why on earth a few hundred bytes of data were worth all of this.It’s no big news that human technology has many of us by the scruff of the neck. Our machines and algorithms serve us, but we serve them too. With its duplicitous nature, social media provides connectivity and opportunity with one hand while it drains our attention and resources with the other. You pay for every Facebook post, Instagram story, and tweet with your own neural activity and investment in hardware and energy.We keep inventing more of such hidden burdens. Crypto enthusiasts expound on the democratic possibilities of decentralized, secure data and currencies derived from blockchain technologies. Yet those technologies can be voraciously resource hungry; it’s inherent to how they work. Other dubious inventions, like non-fungible tokens, rely on those same structures, and machine learning and streaming services consume energy resources as well. Some applications are profoundly useful, yet many appear utterly frivolous for a civilization teetering on the brink of planetary disaster brought on by unthinking resource use.The motherboard of a Sony PlayStation. As gaming becomes more realistic, ever larger amounts of energy are needed to power the world’s players. (Courtesy of Evan-Amos/public domain.)PPT|High resolutionPart of the energetic overhead for all those activities originates in the fundamentals of how we handle information. A modern microprocessor features tens of billions of transistors—structures that represent an extreme reduction of local entropy, which takes a lot of work to accomplish. A much-cited study from back in 2002 introduced the phrase “the 1.7 kilogram microchip,” which references the approximate mass of hydrocarbon fuel and chemicals then required to assemble a single DRAM chip a mere 2 grams in mass. Fabrication also required 32 kilograms of water and about 700 grams of elemental gases.11. E. D. Williams, R. U. Ayres, M. Heller, Environ. Sci. Technol. 36, 5504 (2002). https://doi.org/10.1021/es025643oOf course, the actual running of digital computation is getting more efficient over time. Some improvements come from greater miniaturization; others come from a trend to hardware specialization rather than generalization. The catch is that the tasks we give devices are growing exponentially. Take the example of deep-learning systems: A 2019 study showed how training an all-bells-and-whistles version of the Transformer natural-language processing model, working with over 210 million parameters, can gulp down an amount of energy equivalent to the emission of more than 284 metric tons of carbon dioxide, about the same as the lifetime emissions of five gasoline automobiles.22. E. Strubell, A. Ganesh, A. McCallum, in Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, A. Korhonen, D. Traum, L. Màrquez, eds., Association for Computational Linguistics (2019), p. 3645.An investigation of global data in 2011 found a two-decade trend of about 60% growth per annum in our species’ total computing capacity. That outpaced what continues to be a roughly 20–30% annual growth in data-storage capacity.33. M. Hilbert, P. López, Science 332, 60 (2011). https://doi.org/10.1126/science.1200970 It’s unclear which growth drives which, but perhaps sheer necessity is contributing to computing growth. Still, it’s easy to see that a large proportion of our informational world—including reams of mundane financial data, social media posts of lunchtime sandwiches, and promulgations of false information—has questionable importance for the survival of our species. We don’t really know what the total semantic quality is of the more than 2.5 quintillion bytes of data generated each day by our civilization. Consequently, we wind up expending ever more effort to find benefits.One projection suggests that by 2040 computing will necessitate more energy than the world currently produces.44. Semiconductor Industry Association, Semiconductor Research Corp, Rebooting the IT Revolution: A Call to Action (September 2015). Simultaneously, the total “anthropogenic mass”—all of the matter embedded in inanimate solid objects made by humans—is estimated to already exceed the total biomass.55. E. Elhacham et al., Nature 588, 442 (2020). https://doi.org/10.1038/s41586-020-3010-5The implications of such ideas are both fascinating and concerning. We know that if the resources demanded by our global civilization are not balanced against their environmental impacts, we’ll suffer. At the same time, the vast, externalized informational world that we generate and sustain—an entity that I have dubbed the “dataome” in my 2021 book The Ascent of Information: Books, Bits, Genes, Machines, and Life’s Unending Algorithm—has helped make us one of the most successful and sophisticated species Earth has ever seen. We’ve engineered an astonishing amplification of biological traits by off-loading memory, communication, and problem-solving to other places, outside of our cells and genes.Maybe we can innovate our way out of informational meltdown. Some people pin (perhaps unrealistic) hopes to the realization of more generalized quantum computing. But while qubits use little energy to compute, their environmental conditions require significant power. As of 2015 the hardware of a D-Wave Systems machine consumed about 25 kilowatts of power, much of which was used to maintain refrigeration.66. J. Hsu, “How much power will quantum computing need?,” IEEE Spectrum, 5 October 2015. It’s still unclear how that will scale further. But no matter what, the infrastructure and exponential growth of data storage and retrieval required will remain a burden.Humans may have catalyzed the rise of a dataome and a world increasingly structured and restructured in service of information, but it’s not obvious that the extraordinary benefits we enjoy will continue to outweigh the burdens. The big question is where that problem takes us. Explaining biological evolution has benefited from the concept of the selfish gene, whose ability to propagate relies not on the advantage it bestows but on its ability to enhance its own transmission. The dataome suggests that those resource-seeking informational forms can spill like a tsunami into other domains and follow thermodynamic imperatives that are indifferent to parochial human needs, dissipating energy until our planet’s contents are once again in equilibrium with the rest of a cold cosmos.ReferencesSection:ChooseTop of pageReferences <<1. E. D. Williams, R. U. Ayres, M. Heller, Environ. Sci. Technol. 36, 5504 (2002). https://doi.org/10.1021/es025643o, Google ScholarCrossref2. E. Strubell, A. Ganesh, A. McCallum, in Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, A. Korhonen, D. Traum, L. Màrquez, eds., Association for Computational Linguistics (2019), p. 3645. Google ScholarCrossref3. M. Hilbert, P. López, Science 332, 60 (2011). https://doi.org/10.1126/science.1200970, Google ScholarCrossref4. Semiconductor Industry Association, Semiconductor Research Corp, Rebooting the IT Revolution: A Call to Action (September 2015). Google Scholar5. E. Elhacham et al., Nature 588, 442 (2020). https://doi.org/10.1038/s41586-020-3010-5, Google ScholarCrossref6. J. Hsu, “How much power will quantum computing need?,” IEEE Spectrum, 5 October 2015. Google Scholar© 2022 American Institute of Physics.

We present a video of a simulation showing the expansion front of a technological species settling a Milky Way-like galaxy, created using the model described in Carroll-Nellenback et al. It illustrates how even very conservative rates of settlement ship launches and ship ranges can quickly lead to a galaxy endemic with technology, and how the rotational and peculiar motions of stars contributes to the expansion. This video confirms and validates previous work showing that the centers of galaxies are promising search directions for SETI.

Abstract First contact with another civilization, or simply another intelligence of some kind, will likely be quite different depending on whether that intelligence is more or less advanced than ourselves. If we assume that the lifetime distribution of intelligences follows an approximately exponential distribution, one might naively assume that the pile-up of short-lived entities dominates any detection or contact scenario. However, it is argued here that the probability of contact is proportional to the age of said intelligence (or possibly stronger), which introduces a selection effect. We demonstrate that detected intelligences will have a mean age twice that of the underlying (detected + undetected) population, using the exponential model. We find that our first contact will most likely be with an older intelligence, provided that the maximum allowed mean lifetime of the intelligence population, τmax, is ≥ e times larger than our own. Older intelligences may be rare but they disproportionately contribute to first contacts, introducing what we call a ‘contact inequality’, analogous to wealth inequality. This reasoning formalizes intuitional arguments and highlights that first contact would likely be one-sided, with ramifications for how we approach SETI.

The circumstellar habitable zone and its various refinements serves as a useful entry point for discussing the potential for a planet to generate and sustain life. But little attention is paid to the quality of available energy in the form of stellar photons for phototrophic (e.g., photosynthetic) life. This short paper discusses the application of the concept of exergy to exoplanetary environments and the evaluation of the maximum efficiency of energy use, or maximum work obtainable from electromagnetic radiation. Hotter stars provide temperate planets with higher maximum obtainable work with higher efficiency than cool stars, and cool planets provide higher efficiency of radiation conversion from the same stellar photons than hot planets. These statements are independent of the details of any photochemical and biochemical mechanisms and could produce systematic differences in planetary habitability, especially at the extremes of maximal or minimal biospheres, or at critical ecological tipping points. Photoautotrophic biospheres on habitable planets around M-dwarf stars may be doubly disadvantaged by lower fluxes of photosynthetically active photons, and lower exergy with lower energy conversion efficiency.

Planetary rotation rate has a significant effect on atmospheric circulation, where the strength of the Coriolis effect in part determines the efficiency of latitudinal heat transport, altering cloud distributions, surface temperatures, and precipitation patterns. In this study, we use the ROCKE-3D dynamic ocean general circulation model to study the effects of slow rotations and increased insolations on the "fractional habitability" and silicate weathering rate of an Earth-like world. Defining the fractional habitability f(h) to be the percentage of a planet's surface that falls in the 0 <= T <= 100 degrees C temperature regime, we find a moderate increase in f(h) with a 10% and 20% increase in insolation and a possible maximum in f(h) at sidereal day lengths between 8 and 32 times that of the modern Earth. By tracking precipitation and runoff, we further determine that there is a rotational regime centered on a 4 day period in which the silicate weathering rate is maximized and is particularly strongly peaked at higher overall insolations. Because of weathering's integral role in the long-term carbonate-silicate cycle, we suggest that climate stability may be strongly affected by the anticipated rotational evolution of temperate terrestrial-type worlds and should be considered a major factor in their study. In light of our results, we argue that planetary rotation period is an important factor to consider when determining the habitability of terrestrial worlds.

Many parameters that influence the habitability of a given exoplanet can not be measured. We discuss how can contextual knowledge on exoplanet population and planet formation be combined with uncertain information on individual planets. We review key questions that must be addressed to improve the predictive power of planet formation models.

We model the settlement of the Galaxy by space-faring civilizations in order to address issues related to the Fermi Paradox. We are motivated to explore the problem in a way that avoids assumptions about the agency (i.e., questions of intent and motivation) of any exo-civilization seeking to settle other planetary systems. We begin by considering the speed of an advancing settlement front to determine if the Galaxy can become inhabited with space-faring civilizations on timescales shorter than its age. Our models for the front speed include the directed settlement of nearby settleable systems through the launching of probes with a finite velocity and range. We also include the effect of stellar motions on the long-term behavior of the settlement front which adds a diffusive component to its advance. As part of our model we also consider that only a fraction, f , of planets will have conditions amenable to settlement by the space-faring civilization. The results of these models demonstrate that the Milky Way can be readily filled-in with settled stellar systems under conservative assumptions about interstellar spacecraft velocities and launch rates. We then move on to consider the question of the Galactic steady state achieved in terms of the fraction X of settled planets. We do this by considering the effect of finite settlement civilization lifetimes on the steady states. We find a range of parameters for which 0 < X < 1, i.e., the Galaxy supports a population of interstellar space-faring civilizations even though some settleable systems are uninhabited. In addition we find that statistical fluctuations can produce local overabundances of settleable worlds. These generate long-lived clusters of settled systems immersed in large regions that remain unsettled. Both results point to ways in which Earth might remain unvisited in the midst of an inhabited galaxy. Finally we consider how our results can be combined with the finite horizon for evidence of previous settlements in Earth’s geologic record. Using our steady-state model we constrain the probabilities for an Earth visit by a settling civilization before a given time horizon. These results break the link between Hart’s famous “Fact A” (no interstellar visitors on Earth now) and the conclusion that humans must, therefore, be the only technological civilization in the Galaxy. Explicitly, our solutions admit situations where our current circumstances are consistent with an otherwise settled, steady-state galaxy.