The capacity to bear witness to events and realities is no longer exclusive to humans. In an increasingly interconnected and expansive world, humans are coming to terms with coexisting alongside nonhuman entities capable of knowledge production and witnessing.

The human sensorium alone can no longer register events and crises of vast, planetary, and often imperceptible scales. In contrast, data systems, artificial intelligence, and natural ecosystems possess the ability to ‘read’ these occurrences and engage in acts of worldmaking.

In Nonhuman Witnessing: War, Data, and Ecology after the End of the World², Michael Richardson examines the imperative expansion of the concept of “witnessing” beyond human capacities. This approach serves as a critical framework for understanding systemic and interconnected crises embedded in climate change, technological advancement, and warfare. The book investigates key sites such as nuclear testing, Indigenous territories, autonomous drones, deepfakes, artificial intelligence, and algorithm-driven investigative tools to illuminate diverse perspectives on the ‘Pluriverse’ and strategies for making sense of its complexities³.
 

As we face multiple global crises, what happens when we move beyond a human-centered view of the world? In this conversation, Michael Richardson explores these fundamental questions.

Philip di Salvo: The first topic I would like to discuss is the foundational notion on which your book is built: the expansion of the capacity for witnessing beyond the human sphere. In the book, this concept is closely associated with technological actants such as drones and artificial intelligence. At the same time, significant sections of text delve into the environmental catastrophe that lies at the heart of contemporary challenges. I would be interested in hearing more about this idea— what it entails and how it links technological advancements with the climate crisis.

Michael Richardson: There are a few key impetuses or motivations for the book, including the proposition that we need to rethink witnessing and its location within the human. One of these is certainly the attempt to reckon with the scale of the climate crisis that humans have produced for this planet. If witnessing is indeed a crucial ethical, political, social, cultural, and legal function that elevates certain events to a level of significance where they demand action, then placing the human witness alongside the planetary scale of ecological destruction—or even a regional or localized scale—often reveals scales that exceed human capacity to sense and know. This raises the question: how does a human bear witness or engage in witnessing a crisis and forms of violence that are, on the one hand, broad-reaching and enduring but, on the other hand, difficult to perceive in specific and localized ways? We depend on a variety of abstractions to make sense of the climate crisis. It seemed to me that we needed to push the possibilities of this mode of engagement with the world that we call ‘witnessing’, which has a long and important history across different cultures and contexts. We now need to think about how we might refigure this concept in response to a crisis of such scale. I wanted to explore how we might conceptualize forms of knowing and knowledge-making that are not located to the human.

One potential critique of this position might argue that we already have other ways of discussing knowledge-making and other ways of doing politics. I think that’s true. But at the same time, witnessing has this significant and intriguing cross-cutting quality—it appears in science, religion, media, politics, and journalism. But can’t we rethink what it means? Across academia, scholarship, activism, art, and other fields, there has been a growing recognition that the human is not the locus of all-knowing, not the center of the planet, or even of localized experience—let alone the center of the universe. Why can’t we apply that same logic to knowledge-making concepts that have been so significant for the human and use it to reappraise our relationship to the climate crisis and beyond?”

The book also made me think about the idea of opacity, which recurs throughout its pages. It also brought to mind James Bridle’s work, particularly The New Dark Age⁴, and the difficulty of grasping the dynamics of power or even the dynamics of what can be known. As you mentioned earlier, this idea seems to be part of your discourse on witnessing—how human ways of witnessing and making sense may not be sufficient to fully grasp the complexity of both technology and the climate crisis.

All of this ties into the question of knowability—what is knowable, through what kinds of frames, and how we might shift our points of reference or expand the grounds upon which knowledge can be produced. James Bridle has been a really important thinker in this space, especially around technology with New Dark Age and also with their more recent work on forms of intelligence beyond the human. In this context, I think of their work in unpicking the mysticism surrounding computation. While Bridle is not the only person engaging in this work, they do so in an especially accessible and thoughtful way, moving across a range of different domains. I think the question of what is visible, knowable, and legible to humans—if that remains our way of determining what matters and what we consider significant—will only lead us further down the same path we’ve been following. That’s precisely the logic that brought us here. We need these alternative ways of thinking and approaching the question of opacity, which I find especially intriguing because, on the one hand, opacity is often presented to us as a problem. We could think here about the black boxes of technology, the opacity of machine learning algorithms, and so on. I think that’s true—these are opaque technologies, and within this opacity are elements we should push back against and question. For instance, what does it mean to have significant world-making technologies that are opaque to human understanding at a certain level?

But on the other hand, that doesn’t mean opacity itself is the problem. The real question is: what purpose does opacity serve? Does it enable a kind of justice that allows multiple modes of being to coexist? Or is it an opacity that elevates certain forms of power over others, enabling control and oppression in unjust or violent ways? My thinking around opacity in the book has been deeply influenced by Édouard Glissant, the great Caribbean thinker⁵, who was concerned with reworking opacity away from being a problem that halts communication or a problem that needs to be resolved by making ourselves transparent to one another by making, for instance, all cultures, aesthetics, or modes of being fully legible to each other. Instead, he argued for a “right to opacity,” proposing that coexistence founded on a willingness to accept the opacity we hold in relation to one another refigures the ground of politics in interesting ways.

This approach rejects a flattening universalism, as the transparency that Glissant critiques is precisely the transparency of a hegemonic Western form of knowing. This mode of knowing, arising in his case from colonialism—particularly plantation colonialism—is deeply tied to the Enlightenment, its epistemologies, and the emergence of modern science. It is premised on the belief that we can always look more closely and always make the world more transparent and knowable. Glissant challenges this by asking: what if we truly need recognition of necessary opacity? That said, opacity is not inherently virtuous simply because this argument exists.

Opacity can serve lots of masters, and it has done so. The point is not to uncritically elevate all forms of opacity but to consider its role carefully.

When we think about the production of knowledge, the making of worlds, and the refiguring of our lives on this fragile planet—or at least fragile for human life, as the planet itself will likely endure in the long run; it is we who must worry—we need a greater willingness to accept opacity as a foundation for plural existence.

Thinking about the climate again, we’ve just seen images from Spain, Italy, and now Los Angeles, where the evidence of the problem is undoubtedly clear. There’s a part in the book where you discuss the ecological scales or the multi-historical processes behind what we are witnessing. Speaking as a journalist, or as a journalism scholar, and also referring to Joanna Zylinska’s book on photography at the end of the world⁶, how are we going to report on such a piece of news? How do we make news of the end of the world? Also, beyond journalism, of course.

I think journalism is very bound by its particular rules and genres, and in many places, it is still bound by the notion of objectivity, as well as the necessary grounding in specific kinds of evidence and ways of narrating what has taken place. This is even before we get to the pernicious effects of the need for balance in certain journalistic contexts. For instance, if you talk about the damage and destruction caused by climate change, you must somehow temper that with different perspectives. My observation is that climate journalism has shifted away from this type of coverage and now positions these things as facts.

I think journalism, activism, art, and other domains trying to make the climate crisis legible, all represent spaces where stepping outside the human frame could help us understand and uncover the ecological roots, implications, and impacts. This also extends to writing, talking, or producing imagery about the crisis in ways that don’t reduce it to what it means for human existence, or that assume the nonhuman or ecological concerns are only of interest in narrow ways”.

“I don’t really know, unfortunately, how to operationalize the notion of nonhuman witnessing in those types of contexts, or rather, I don’t know them. I do think, however, that in some ways, this doesn’t necessarily need to be an absolutely radical gesture; it can be about where we shift our focus—what we position as significant, or where we position knowing, and what the lens through which we examine things might be. Do we tell the story of something through the gaze of the journalist, or do we tell the story through the impact on a landscape or the destruction of a complex local ecology? My thinking about nonhuman witnessing is not to banish the human, of course, but rather to place a new kind of injunction upon the human—to attend to the nonhuman and to recognize that there is no human without the nonhuman. In this sense, we are always both human and nonhuman at the same time.

Moving to the media and technology side of the topic, the book also discusses contemporary media consumption, which is linked to how humans contribute to accelerating the climate crisis. Of course, this is also connected to the enormous ecological costs of AI and crypto.

“I understand media in my work in a very broad, maximalist sense, similar to the approach John Durham Peters takes, particularly in his book The Marvelous Clouds⁷. I position my reasoning within the materiality of media but also within theoretical fields that explore the expansiveness of media. I’m not talking about ‘the media’ in the sense of journalists or Netflix—though that is part of it—but rather in a much broader sense: anything that performs the work of mediation. I believe it’s a helpful way of addressing questions about interrelation and relationality. Recognizing processes of mediation as fundamental ways the world is composed—such as how chemical reactions occur through media—can be valuable. The notion that media consumption and mediation processes are expanding, accelerating, and contributing to the planetary ecological crisis is one I borrow from Sean Cubitt in his book Finite Media⁸, where he points out that media is finite because the planet's resources are finite”.

Not only are the planet’s resources finite, but the entropy of the universe means that, while you can turn matter into energy, you can’t turn energy back into matter. As we engage in processes of mediation that entail the transfer of matter into energy—such as burning coal to get electricity to power our infrastructures and communication networks—we are running down the finite balance of the planet and its capacity to sustain this. Yes, we can get energy from the sun and other sources, but none are tapping into something infinite. They all depend on the consumption of finite resources. So, when we look at where the climate impacts are coming from now, they are increasingly linked to media production, particularly artificial intelligence, as you mentioned. The degree to which this shift has occurred in just a few short years is truly astonishing. I began writing this book—though it has longer roots—in 2019 or 2020, right amid the machine learning era of computation, but not yet in this generative AI era. Discussions about energy-intensive, resource-heavy, and ecologically destructive computation were already present, but even then, I don’t think we could have imagined how much those issues would escalate in magnitude. Now, we’re at a point where old, faulty nuclear power plants like the Three Mile Island plant are being reopened to power-generative AI, which—while not dismissing the technology entirely—still doesn’t seem to offer many clear or useful applications for people. A very strange thing is taking place, where the finitude of media and its impact on the planet is accelerating precisely at the wrong time. In the book, I suggest that we live in a period of interlocking crises. There are many, but the three I identify are increasingly autonomous warfare, algorithmic enclosure, and ecological catastrophe. With the explosion of generative AI, we can see how these three crises are entangled.

These three crises are explicitly entangled in a company that, ironically, is named after the Amazon. 

Exactly. And we’re seeing a war in Gaza, which is being massively accelerated and intensified by algorithmic technologies, causing considerable ecological destruction alongside the terrible human cost and the devastation of homes, infrastructure, hospitals, universities, and more. Meanwhile, militaries are also rushing to tap into the generative AI hype machine, trying to grasp it. So, we have these intersecting crises that are only intensifying. It’s astonishing—but, frankly, unsurprising—to see companies like Google, Amazon, and Microsoft basically stepping back from, if not abandoning entirely, their net-zero and sustainability objectives because they no longer align with this new form of technological power grab in the form of generative AI. This is precisely how these crises interlock. And just recently, Eric Schmidt, former head of Alphabet and now chair of the U.S. Defense Innovation Board, suggested that the 1.5°C increase threshold had essentially been surpassed and that the situation was irreversible⁹. He indicated that generative AI had sealed the fate of that target, leaving only the hope that generative AI would somehow solve the problem. It's like you’ve set your house on fire by spilling gasoline, but you don’t have any water to put it out. So instead, you pour more gasoline on the fire, hoping it will somehow ‘get with the program’ and put itself out, fixing your house in the process. It’s completely absurd, right? But in some ways, this is where we find ourselves. To reckon with what’s going on, we need to engage with both the very human notions of witnessing and understanding violence, alongside other ways of making sense of the situation.

The necessity, I think, is to reckon with nonhuman agencies and their knowledge-making capacities in a way that doesn’t reduce them to mere evidence or input data for human reckoning. Instead, we need to ask: what is the process of knowledge-making involved in creating a foundation model for generative AI, for example? And how might we consider the interrelations between those foundation models, the model-making process, or the GPT query, and their ecological entanglements, costs, energy use, impacts across supply chains, and so on? All of this, of course, always involves humans at different levels—almost always—but often in a state of subservience or lacking agency, at least in relation to those larger systems.

Another thing I want to touch on with you is OSINT investigations, which are becoming one of the preferred methods for witnessing conflicts, as well as documenting climate disasters and other evidence of crises, in journalism and beyond. These investigations often combine human and nonhuman witnessing—for example, satellite imagery paired with on-the-ground reporting or personal accounts of events. I find this combination extremely interesting. It seems to exemplify what we discussed earlier about the co-presence of human and nonhuman actors and actants in these discourses.

I think OSINT investigations are such a good example of engaging with nonhuman agencies, nonhuman ways of sensing, and the nonhuman registration of violence—whether in landscapes, urban structures, or other environments. Thinking of Forensic Architecture work¹⁰, I see those as instances of trying to grapple in a meaningful, practical, and operative sense with nonhuman witnessing. In my conversations with people in that community—some with folks from Forensic Architecture and at Airwars in the UK¹¹—there seems to be a real recognition of the necessity of working with the nonhuman and engaging with it in various ways. They don’t necessarily frame it in those exact terms, but the way these issues are discussed feels distinct from, say, someone positioning themselves as the sole investigative authority or privileged witness of what’s taking place. It’s completely different.

I've been very influenced and inspired by Forensic Architecture’s work and, you know, by Eyal Weizman’s writings, including the recent book with Matthew Fuller¹². I’ve been lucky enough to have some conversations with Eyal over the years, and I’ve found those super interesting. I think our point of difference—though I don't think it’s really a disagreement—might be a question of emphasis or interest in action. Forensic Architecture, and I think many other OSINT organizations, are deeply focused on entering their work into the public forum and have their findings presented with the assertion that the work must clearly convey what has been discovered and the actions being taken based on those discoveries. The importance of opening their work to debate, ensuring it is contestable, and making the evidence available is emphasized. Without fostering that contestation, it is believed that a true engagement in truth-making as a collective social enterprise cannot occur. I don’t disagree with anything of this view, but my focus has largely been on what comes before all of that. That’s why I’m more interested in nonhuman witnessing as a process—witnessing in the gerund form rather than the witness as a figure, the forensic evidence as an outcome, or the findings of a report. Of course, those are all incredibly important in various ways, but my book and much of my thinking have centered more on what makes up all of that before it takes that particular form.

Could all of this not also apply to artificial intelligence, undoubtedly the form of technological and nonhuman way of knowing we are increasingly getting used to live with? AI is also being used to assist OSINT investigations.

In some cases, there’s not always this awareness, but in many instances, there is a real understanding of the pitfalls of using artificial intelligence, such as issues of bias and discrimination. People are aware of the problems of inaccuracy, hallucination, and other issues. They recognize these problems but also acknowledge that flaws exist in all sorts of tools. They believe we can apply technologies in different ways and at various scales. We might not achieve perfect outcomes, but we can experiment with them. One way to mitigate these problems might be to make our work transparent or to make the models and tools we use open source, allowing people to interrogate, improve, and see how they arrive at their conclusions. Again, this doesn’t guarantee perfect outcomes, but many of these organizations are not working within the domain of perfection—they operate within a domain of strategic pragmatism. I have a lot of sympathy for that because, in some ways, that's how theory should operate as well—strategically and pragmatically. In fact, I mention in the book that, for me, nonhuman witnessing is a kind of strategic concept. It’s a strategic gesture to argue that we need new ways of thinking about knowledge-making within this context of crisis. We must shift the ground of witnessing. By privileging the human as the center of ethical and political responsibility making, we exclude so much. It’s not that we necessarily want nonhuman witnessing as a concept to endure forever, but rather that we need to shift the conversation. Perhaps, eventually, we can move beyond the term or find new ways of making knowledge that doesn't rely on such interventionist concepts. But until then, we may need to continue working with them. In my work, as I wrote the book, I became not only concerned with the nonhuman aspect of witnessing but also with the problem of the human. It’s not just about pushing beyond the human; it’s also about questioning what we mean when we talk about the human witness. In this respect, I learned a great deal from reading black studies work from the United States, reading Édouard Glissant, who I mentioned earlier, and reading work by First Nations scholars both in Turtle Island, North America, and in Australia, where I’m from. This has been really important in thinking about different articulations of the relation between the human and the world the human exists within, and also recognizing the violence that has been inherent to the category of the human.

Are we, by engaging with the nonhuman, for instance via artificial intelligence, also questioning the notion of ‘the human’ in its hegemonic characterizations?

One problem with the human is that many people worldwide have been historically refused that category. In the case of the United States, for example, during slavery, the notion that enslaved Black people were human was explicitly rejected. And when I say that, I mean that quite explicitly around legal categories, such as standing before the law and rights within courts. The same is true in many other settler societies, where Indigenous people, enslaved people, nonwhite people in general, or even unpropertied people were not able to have standing before the courts. In a very real sense, they were denied the same kind of humanity as others. And, you know, the era of human rights since the 1940s has not reversed that. In fact, in some ways, the era of human rights has actually further instantiated or solidified the category of the human as the necessary locus of rights, political action, and so on. So, I wanted to say, well, okay, yes, we need to take seriously what exists outside or beyond the human. But we also need to question why it is that we have elevated the human into this category of witness. Because the figure of the witness as a human witness has, I think, historically been really associated with what Sylvia Wynter calls “the man with a capital M.” Her argument is more complicated than I’ll gloss over here, but she essentially says that there is this figure of man, a white European Enlightenment figure, that over-represents itself as ‘the human’. The figure we are given as human is not, for example, the Darug First Nations person from the area around where I live in Sydney. It’s not the enslaved Ghanaian. It’s not the person brought from the Ivory Coast. That’s not the figure through which the human is refracted. It’s a very different figure. And I think once you recognize that, you can understand why there is a certain hostility to human rights regimes in various parts of the world. While some of this hostility is disingenuous—like China’s opposition to human rights, partly because it wants to continue oppressing its people—there is also a genuine rejection of the notion of a particular kind of figure of man: male, white, European, and presented as standing in for the universal. If that is the witness, then that figure of man, the witness, is the figure of the witness in science, before the law, in politics, and often in media and journalism. Many non-white, non-male individuals take up this figure of what a witness is because it’s what they’re given as a possibility. So, reckoning with the nonhuman in all these crises is also about reckoning with the human and the status of human knowing in itself.

About the authors:

Philip Di Salvo is a researcher and lecturer at the School of Humanities and Social Sciences, University of St. Gallen (HSG), Switzerland. His research interests include investigative journalism, internet surveillance, the intersection of journalism and hacking, and black box technologies. Previously, he served as a Visiting Fellow in the Department of Media and Communications at the London School of Economics and Political Science (LSE) from 2021 to 2022. From 2012 to 2021, he held various research and teaching roles at the Institute of Media and Journalism, Università della Svizzera italiana (USI). In the summer of 2024, Philip was a Visiting Fellow at Harvard University. Between 2018 and 2020, he also taught as a Lecturer at NABA – New Academy of Fine Arts in Milan, Italy. As a freelance journalist, Philip has contributed to publications such as Wired, Motherboard/Vice, and Esquire, covering the societal impacts of technology. Additionally, he hosts a monthly technology-focused radio show for Milan-based Radio Raheem. Philip is the author of two books: Leaks. Whistleblowing e hacking nell’età senza segreti (LUISS University Press, Rome, 2019) and Digital Whistleblowing Platforms in Journalism. Encrypting Leaks (Palgrave Macmillan, London, 2020). He also serves on the board of DIG Festival, an international investigative journalism event held in Italy.

Michael Richardson is an Associate Professor of Media at UNSW Sydney, where he co-directs the Media Futures Hub and Autonomous Media Lab, and an Associate Investigator with the ARC Centre of Excellence on Automated Decision-Making + Society. Drawing on a transdisciplinary background in media studies, cultural studies, literature, and international relations, his research examines technology, violence, and affect in war, security, and surveillance. Michael is the author of the books Gestures of Testimony: Torture, Trauma, and Affect in Literature (Bloomsbury 2016) and Nonhuman Witnessing: War, Data, and Ecology after the End of the World (Duke UP, 2024). His research also appears in edited collections and leading academic journals such as Theory, Culture & Society, New Media & Society, Continuum, Environmental Humanities, Cultural Studies, and Media, Culture and Society. Michael also writes for non-academic outlets, such as ABC News, The Conversation, and Sydney Review of Books. Michael held an ARC Discovery Early Career Researcher Award and has been the recipient of internal and external grants. An experienced media commentator, Michael's expertise extends across a range of topics at the nexus of media, technology and culture, including drone warfare, surveillance, algorithms, cultural trauma, affect and emotion, political violence and torture. He is available for comment on these and related issues and contacted by email or via the UNSW Newsroom.

Footnotes

1  Header Image: “Morenci Mine #2,” Clifton, Arizona, Edward Burtynsky, 2012. © Edward Burtynsky, courtesy Sundaram Tagore Galleries, Singapore / Nicholas Metivier Gallery, Toronto

2 Richardson, Michael. Nonhuman Witnessing: War, Data, and Ecology after the End of the World. Duke University Press Books, 2024.

3 Escobar, Arturo. Pluriversal Politics: The Real and the Possible. Durham: Combined Academic Publ., 2020.

4 Bridle, James. New Dark Age: Technology and the End of the Future. New Edition. London Brooklyn, NY: Verso Books, 2023.

5 https://en.wikipedia.org/wiki/%C3%89douard_Glissant

6 Zylinska, Joanna. Nonhuman Photography. Cambridge, Massachusetts London, England: MIT Press, 2024.

7 Peters, John Durham. The Marvelous Clouds: Toward a Philosophy of Elemental Media. Chicago, Illinois, London: University of Chicago Press, 2015.

8 Cubitt, Sean. Finite Media: Environmental Implications of Digital Technologies. Illustrated Edition. Durham, N.C. London: Combined Academic Publ., 2016.

9 https://www.theregister.com/2024/10/08/eric_schmidt_speech/

10 https://forensic-architecture.org/

11 https://airwars.org/

10  Fuller, Matthew, und Eyal Weizman. Investigative Aesthetics: Conflicts and Commons in the Politics of Truth. 1. Aufl. London ; New York: Verso, 2021.