Taghrid Bou Merhi: Is Analog Memory Fading Before Generated Memory? And Are We Witnessing a Transformation in the Structure of the Self or Merely in Its Tools?

Illustration: AI

Since the French philosopher Henri Bergson introduced his conception of memory as a living duration in which temporal layers interpenetrate within human consciousness, memory has no longer been understood as a mere reservoir of past images; it has come to be seen as a dynamic temporal fabric that permeates the present and continuously reshapes it. Bergson distinguished between mechanical memory, linked to habit, and pure memory, rooted in the depth of lived experience, where a moment is not retrieved as it was, but reawakened in a new form within the context of the present. This philosophical framework provides a crucial entry point into the contemporary dilemma we face in the age of generative artificial intelligence. The issue is no longer limited to expanding storage capacity or accelerating retrieval, but concerns a transformation in the very nature of memory itself, a shift from lived trace to generative model.

Since the French philosopher Henri Bergson introduced his conception of memory as a living duration in which temporal layers interpenetrate within human consciousness, memory has no longer been understood as a mere reservoir of past images; it has come to be seen as a dynamic temporal fabric that permeates the present and continuously reshapes it. Bergson distinguished between mechanical memory, linked to habit, and pure memory, rooted in the depth of lived experience, where a moment is not retrieved as it was, but reawakened in a new form within the context of the present. This philosophical framework provides a crucial entry point into the contemporary dilemma we face in the age of generative artificial intelligence. The issue is no longer limited to expanding storage capacity or accelerating retrieval, but concerns a transformation in the very nature of memory itself, a shift from lived trace to generative model.

Within the current digital landscape, memory is no longer solely an organic function embedded in the human brain. It has become part of a vast algorithmic system capable of analyzing data, detecting patterns, and producing plausible configurations of what may be interpreted as “the past.” Here a fundamental question arises: when memory shifts from retrieval to suggestion, from an imprint rooted in bodily experience to a probabilistic structure open to recomposition, does its meaning remain intact? Or are we facing a profound displacement affecting the structure of perception and identity alike?

Analog memory, in its anthropological sense, has always been tied to the body and to lived time. It emerged from sensation, was anchored in emotion, and took shape through language as an extension of experience. Human beings do not remember in a neutral manner; they reorganize the past according to the conditions of the present. Feelings intertwine with details, and events are reconstructed within a singular subjective framework. Memory was never an exact replica of what occurred; it was a living reconfiguration in which forgetting interacted with preservation, and absence played a role as decisive as presence.

With the rise of cloud computing and generative systems, a new phase has emerged that surpasses the earlier notion of digital storage. In the initial stages of digitization, data were converted into binary code, preserved electronically, and retrieved with precision. Generative artificial intelligence, however, does not merely recover stored material; it produces new content derived from prior patterns and offers persuasive possibilities that may lack any lived origin. Memory thus becomes a probabilistic space capable of generating images, texts, or voices that resemble the past without extending from a specific experience.

This shift leads to the erosion of memory’s analog character, meaning the weakening of the direct bond between experience and its imprint. In the bodily model, recollection was linked to neural interconnections formed through lived events and shaped by sensory and emotional factors. In the generative model, memory transforms into a network of mathematical probabilities, constructed through the analysis of vast datasets and reorganized according to statistical weighting. The distinction between the two models exceeds differences in tools; it concerns the nature of the relationship between the self and its past.

From an anthropological perspective, memory has long functioned as a foundation of collective identity. Rituals, narratives, myths, and oral archives ensured communal continuity across time. Memory operated as a social practice, retold and reformulated within a specific cultural context. Today, algorithms participate in shaping this memory, proposing to individuals what they might remember through intelligent archives and digital reminders that reorganize the past according to technical logic. Time becomes material for rearrangement rather than a linear extension of lived existence.

This transformation appears clearly in the phenomenon of “suggested memories,” where digital platforms display images or events from previous years, summoning the past at a designated moment. This recall does not arise from inner desire; it is prompted by an algorithm selecting what it deems suitable. The center of agency shifts from the individual to the system. One receives one’s past ready-made, reordered, framed within a visual or textual context that subtly redirects its meaning.

Another prominent sign of analog erosion is the expansion of archiving. Everything is documented, preserved, republished. The paradox is that excessive preservation may diminish value. A moment loses its uniqueness when absorbed into an endless stream of data. Memory once depended on selective emphasis; it now confronts an abundance that unsettles meaning. When every event becomes recordable, the distance that once granted certain moments their significance begins to fade.

The impact deepens at the level of the self. In the analog framework, the self emerged through the accumulation and interpretation of experience. Remembering functioned as a reflective act, enabling renewed understanding of the past. Within the generative environment, the self engages with multiple versions of its own representation. Texts can be produced expressing emotions never lived, images can simulate events that never occurred, and voices can imitate specific tones. This multiplicity may contribute to a fragmentation of identity, where original experience merges with generated representation.

The essential difference lies in reference. Analog memory rests upon a bodily and temporal origin, even when reshaped. Generated memory relies on patterns and data and can produce convincing effects without attachment to a specific event. This displacement raises profound questions concerning truth itself. If a persuasive narrative of an event that never happened can be generated, how can one distinguish between what was lived and what was suggested?

This does not imply that technology eradicates human memory; it reshapes the conditions under which memory operates. Artificial intelligence may be understood as an extension of human imaginative capacity, distinguished by scale, speed, and analytical depth. Yet such extension carries the risk of normalizing generation, rendering produced content self-evident and obscuring the question of origin. When critical awareness diminishes, the capacity to discern experience from representation weakens.

Within this context, intellectual and ethical resistance becomes necessary. Resistance entails reframing the relationship with technology rather than rejecting it. It involves restoring slowness in an era of acceleration, prioritizing direct experience, and cultivating awareness of generative mechanisms. The body, with its sensations and limits, remains a primary reference that cannot be entirely replicated. The experience of pain, joy, or awe belongs to living presence and cannot be reduced to algorithmic processing.

Writing represents one form of such resistance. Through writing, an individual reconnects language to personal experience and inscribes memory with a distinct imprint. In an age of generation, writing becomes a contemplative act that preserves the link between word and body. The aim is not solely the production of unique texts, but the reaffirmation of the bond between the self and its lived past.

The future of memory remains an open question. Generative systems may evolve to a point where the boundaries between remembering and producing grow increasingly indistinct. Nevertheless, human beings retain the ability to sustain critical distance, enabling engagement with technology without surrendering their grounding. Analog memory may recede in visible prominence, yet it persists in art, in embodied encounter, and in experiences that resist reduction to data.

The erosion of the analog does not signify the end of human memory. It indicates a transformation in its structure and function. We inhabit a transitional moment in which modes of memory overlap, the bodily and the digital coexist, and origin and possibility confront one another. The central challenge lies in recognizing this shift, interrogating it thoughtfully, and preserving the distance that prevents complete absorption into the logic of generation.

Memory remains an existential act exceeding its cognitive dimension. It is the means through which a person understands oneself and grants life coherence and meaning. Media may change, tools may advance, and intelligent systems may expand in capability, yet the question of who holds the past and who defines its meaning endures as a fundamentally human concern. It calls for sustained intellectual vigilance and reminds us that lived reality cannot be measured solely by what can be generated, but by what is experienced within the depth of living consciousness.

Lebanese-Brazilian poet and translator Taghrid Bou Merhi