What is lost when knowledge moves through machines rather than bodies?
69066
wp-singular,post-template-default,single,single-post,postid-69066,single-format-standard,wp-theme-bridge,wp-child-theme-WACC-bridge,bridge-core-3.3.5,qodef-qi--no-touch,qi-addons-for-elementor-1.11.2,qode-page-transition-enabled,ajax_fade,page_not_loaded,,qode-title-hidden,qode-smooth-scroll-enabled,qode-child-theme-ver-1.0.0,qode-theme-ver-30.9.1,qode-theme-bridge,qode_header_in_grid,qode-wpml-enabled,wpb-js-composer js-comp-ver-9.0.1,vc_responsive,elementor-default,elementor-kit-41156

What is lost when knowledge moves through machines rather than bodies?

Nathaly Espitia and Denisse Albornoz

This article contains two voices: mine, Nathaly Espitia, speaking from the Cañaveralejo River basin in Cali, Colombia, and Denisse Albornoz, speaking from the shores of the Pacific Ocean in Lima, Peru. Together, we reflect on the transmission and preservation of knowledge and memory through communication and new technologies. I’ll speak first.

In 2017, I was waiting for a plane while listening on a phone, together with the other members of Nois Radio, to the news that, by a very small margin, the “No” vote had won the peace referendum. Colombia had decided to vote against reaching an agreement. In the seats of that small airport, the sadness was palpable. How could a country immersed in an armed conflict that had lasted more than six decades decide not to pursue an elusive peace that had not been achieved through armed force?

That same year, through other mechanisms, the Santos government managed to begin the process. The agreement included an exercise to create an archive and preserve the memory of the conflict – one that, among other things, sought to repair the victims in multiple ways and outside the framework of traditional justice, including by preserving and recounting the memory of the conflict. In this sense, communication through multiple voices and in digital spaces has been vital in seeking to uphold an agreement that has been imperfect, poorly executed and has done little to fulfil its promises to victims.

On June 21, 2026, I was once again sitting, listening to the news that Abelardo de la Espriella had become the new president of Colombia. The proposed approach of his government was to resume a constant and direct armed struggle through the use of weapons. In August 2026, Abelardo de la Espirella appointed a conflict denier as director of the National Center for Historical Memory, the now director of the centre proposed, in 2022, defunding mechanisms such as the Truth Commission, which had been responsible for safeguarding and preserving the conflict archive, including in digital form.

How have communities responded to the power structures that decide what is told and what is preserved in the era of digital spaces, social media, and artificial intelligence? And from which voices, and in what languages, is historical memory – the kind that helps prevent repetition of the conflicts – being told?

Sites of contestation

This question has become more urgent as governments across the region have reduced or reorganized the public institutions responsible for memory, culture, and Indigenous affairs. A few months ago, my colleague Denisse, who lives in Peru and is also writing this article with me, described a visit to the Lugar de la Memoria y la Reconciliación, one of Lima’s memory museums, now facing threats of defunding and closure.

Like in Colombia, in Peru, the struggle over memory is a question of whose bodies can exist safely in public spaces, which territories are protected or made to bear the costs of extractivism, and the languages in which stories and worldviews are imagined and told. The Truth and Reconciliation Commission, established to investigate the political violence that devastated Peru between 1980 and 2000, titled the summary of its final report Hatun Willakuy, Quechua for el gran relato, or “the great story”, as a means to amplify the thousands of indigenous voices that were killed in that period. The Lugar de la Memoria y la Reconciliación was created in part to ensure political violence and terrorism would not also extinguish their place in the nation’s history. Yet the institution itself has become a site of contestation, questioning how the voices lost in this conflict are narrated and remembered.

Outside the halls of a museum, memory lives in how we gather, care, cultivate, create, and share what we know. As the government of Keiko Fujimori took office in Peru, continuing the political legacy marked by the crimes against humanity for which Alberto Fujimori was convicted  – including the forced sterilization of thousands of mostly rural and Indigenous women – I keep thinking of the women who continue to provide nourishment and collective care when the state fails them. In the women who organize comedores populares and ollas comunes; the Kukama women who went to court to defend the Marañón River and its right to flow freely; the seed guardians of the southern Andes who are recovering and sharing native seeds in response to a changing climate; and the Shipibo-Konibo women of Cantagallo, who encode ancestral knowledge into kené, weaving it across textiles and murals and bringing into public view histories and ways of knowing that official institutions have often left out or made difficult to see.

These are all forms of acuerpamiento: bodies gathering and sustaining one another, keeping alive forms of knowledge that have endured beyond the institutions meant to preserve them. As more of these memories are carried into digital spaces – what happens to collective memory as the knowledge we carry is increasingly shaped by new technologies?

AI and our infrastructures of memory

Since the widespread introduction and pervasiveness of AI in the collective imagination, self-organized activists and communities are experimenting with new ways to hold community spaces that preserve memories that have sustained their wellbeing and lives, while assessing what it would mean for AI to become part of this infrastructure. There is an undeniable tension between the possibilities afforded by these technologies and the disconnection, addiction, dependency, and other digital mechanisms that can increasingly separate people from human contact.

When and why use AI? Is it the future? The past is constructed in the present, or, as some Indigenous communities say – and right now I am thinking of the Misak in Colombia – the future lies behind us, and the past is what we are walking toward. Time, then, is not linear. What lies ahead is knowledge: cultural heritage and generational wisdom. What paves the way is what has already happened: the decisions we make and whether what we create is in harmony with nature and the ecosystem. In Indigenous and rural communities, knowledge is carried forward without haste, and decisions that affect the whole community are made collectively.

In the race toward AI, this principle has not necessarily been followed. The idea that we need to inhabit that space – including by developing datasets from Indigenous communities that go beyond language – may serve technology more than the communities themselves. New technologies have helped address specific challenges, such as access to information in Indigenous languages.  At the same time, communicators and activists have questioned the urgency behind their implementation.

A few days ago, I heard a colleague ask, “Why do we have to respond to that call?” Why the urgency regarding our data? Why, if it is not a necessity for us, is it presented as one? Indigenous, Afro-descendant, and Black communities are debating the advantages of entering that space, who keeps their data, for what purposes, for how long and under what frameworks. After protecting their knowledge for decades, they are now being urged to hand it over through a narrative that suggests they cannot progress, innovate, access rights, or be included without embracing this technology.

Across Latin America, AI initiatives are bringing Indigenous languages, many of them transmitted largely through speech, into digital spaces where they have long been poorly represented. In Peru, some efforts focus on language revitalisation. Voces, tecnología en tu lengua is developing translation AI tools from an intercultural perspective with Iskonawa and Shipibo-Konibo communities, while Siminchikkunarayku uses crowdsourcing to build digital text and voice corpora for endangered languages, including Quechua. Other projects, such as Illariy and Quispe Chequea, are using AI to bring news and fact-checking in Quechua, Aymara, and Awajún to wider audiences, including through community radio.

But what happens when an oral tradition has to be adapted to a technological system? The creators of Illariy, for example, found that some Quechua consonants could not be reproduced and struggled to approximate the language’s spoken rhythm. Technical choices will shape how language is represented in digital systems, sometimes in ways that may not fully reflect how it is embodied and passed between generations. What futures are being imagined through these technologies for the many worlds carried through oral tradition?

While these uses of AI emerge from real needs and aspirations, their effects on culture and intergenerational knowledge sharing may only become visible over time. AI is being introduced into activities that have traditionally been sustained through relationships. What happens to the social fabric when some of its core functions, such as building trust, reciprocity, and care, become mediated by algorithms and technology companies shaping what is preserved and heard, and who becomes visible in digital spaces?

Possible futures built on community, pauses and dignified hope

This is the time to think about how these technologies can remain community-owned and responsive to the relationships and forms of knowledge they are meant to support, rather than asking those relationships to adjust to the protocols and logics of digital technologies.

This question has stayed with me since the earthquake in Colombia on 10 August 2026: where is the memory of an event that exposed us to so much pain and violence stored, and how is it communicated? From 7:34 a.m. onward, digital and analogue technologies shaped the experience of the affected cities. In Cali connectivity was intermittent during the first hour. After 8:10 a.m., Comuna 19, the hardest-hit area, lost almost all connection. Big Tech failed us when communication was most urgent, and for many hours, radio was our only way of learning what was happening elsewhere. At the same time, mobile devices preserved memories of those days spent recovering lives and bodies from the rubble. Technology – both analogue and digital – helped establish timelines, document what we wanted to remember, and support rescue efforts. As part of Nois Radio, I was also one of the co-coordinators of a sound brigade, where we used sound technologies and our knowledge to help rescuers work more accurately.

At the various collapsed buildings, volunteers created archives from materials recovered from the rubble. Local digital media outlets took it upon themselves to tell stories beyond those covered by local, departmental, and national authorities, whose narratives often seemed designed to portray government officials as heroes without addressing the complexity of the situation or giving voice to diverse and dissenting perspectives. This experience is a perfect example of why communicators and channels that amplify local information matter.

New technologies can be allies, but not when they homogenize complex local problems and ways of knowing or impose quick, universal solutions. Perhaps the future lies in returning to local technologies and intimate conversations: tools and exchanges that respond to specific needs, remain adaptable, and can still inform broader discussions among communities that see themselves reflected in them.

Historical memory is constructed – and must be constructed – outside institutions, otherwise we risk having only one side of the story presented and preserved. This is where the community’s autonomous, self-managed infrastructure has become vitally important: ensuring that the historical memory of our losses, our grief, our stories and the different ways the earthquake impacted us are told and shared. Above all, it ensures that they are not forgotten – not only because of the speed at which algorithms operate, but also because of the speed with which those in power shift the focus and other tragedies become headlines.

Amid this sense of urgency, communities are pausing for dialogue, debate, and analysis, making decisions at their own pace and going back to where knowledge lives. The same pause is necessary when we think about AI. Beyond using it, what are the conditions for appropriating it? How should communities decide whether, how, and for what purpose it is needed, and to whom their data should be provided? What infrastructures are built and which are eroded in the process?

At this moment, AI infrastructure is not ours. Even open-source developments depend on infrastructure created within the Global North, by powers far removed from our Latin American territories. This infrastructure requires massive capital and reflects a relationship with nature that does not place harmony and life at the centre – unlike the worldview of rural, Indigenous, Afro-descendant, and Black communities. This is a call to decentralize digital technologies and take a breath before acting. Where is our knowledge headed? What relationships are lost when knowledge is transferred to machines rather than between generations? How can we imagine a future that grows from our relationships, the land, and the ways we have learned to care for one another?

Disclosure: The text in this article was written by both authors. AI tools were used to correct grammar and improve readability, after which it underwent another round of human editing to ensure our voices and the heart of our ideas were still present.