Collective memory in the age of algorithms
Algorithms don't just manage our feeds: they are rewriting collective memory, deciding what we remember online and what disappears into the digital noise. Betwe
Collective memory is the way a society remembers itself through shared stories, images, archives, and rituals. Today, this process increasingly passes through algorithms that select, order, and resurface what we see online, transforming the internet into a kind of "global archive" that is both dynamic and unstable.
Introduction: Why It's an Urgent Topic Today
When you open a social media app, search for something on Google, or let a virtual assistant summarize the history of an event for you, you are delegating to an algorithm the choice of which pieces of the past deserve to be remembered. It's not just about your photos or messages, but the memory of entire communities: wars, traumas, achievements, injustices, moments of celebration and crisis.
The digital world has transformed memory from a "box of the past" into a continuous stream of suggested memories, "on this day" notifications, resurfacing posts, and articles re-proposed because they've suddenly gone viral again. With generative artificial intelligence, this stream doesn't just preserve: it reworks, synthesizes, rewrites, and creates new narratives from the traces we leave online.
In this scenario, talking about collective memory means talking about power: who decides what gets archived, what gets put on the front page, what remains on the margins or is silenced. And this is where not only technicians come into play, but also citizens, institutions, educators, and global projects like UNESCO's Memory of the World programme, which tries to protect humanity's documentary heritage in digital form.
What is Collective Memory in the Digital Age?
Collective memory, in the classical sense, is the set of memories, symbols, and narratives that a society builds and passes down through institutions, media, schools, rituals, and culture. In the digital age, as highlighted by contemporary sociological studies on memory, these channels intertwine with global platforms, cloud archives, social media, and search engines that become central infrastructures of shared remembrance.
Today, a significant part of collective memory is built on platforms that archive everything: videos, comments, posts, articles, digitized historical documents. In parallel, thousands of local and global initiatives (Holocaust archives, photographic collections, cooperative documentation projects) use digital tools to make historical memory more accessible.
But collective memory is not just what is preserved: it is also what is highlighted, contextualized, and reinterpreted in the present. And this part today is heavily mediated by algorithms that decide which content to push, which links to suggest, which news to resurface based on criteria of attention, engagement, and profiling. Platforms are not just passive archives, but true algorithmic editors of social memory.
How Are Algorithms Rewriting Collective Memory?
The algorithms that govern social media, search engines, and recommendation systems do not just store information: they select what becomes visible, viral, and memorable. They function as a kind of "invisible archivist" that decides which stories emerge and which are buried deep, often without the user realizing it.
On this topic, we have already shown how algorithms can amplify invisible biases and discrimination, influencing access to opportunities and social justice. The same mechanism acts on collective memory: if the data on which algorithms are trained and function is incomplete or distorted, the resulting narrative of the past will be too.
In the case of AI-fueled misinformation, algorithms often reward polarizing and emotionally strong content, favoring the spread of false or manipulated narratives that end up settling as public "false memory." As analyzed in recent studies on AI-driven misinformation (Harvard Misinformation Review), generative tools are capable of producing false political news almost indistinguishable from real ones, with a direct impact on the perception of recent history and democratic processes.
Towards a Global Archive? Opportunities and Risks
Can we really speak of a global memory archive? In part, yes: thanks to digitization, an enormous quantity of documents, images, testimonies, and works is available online, accessible from anywhere in the world.
However, this "global archive" is anything but neutral or complete. There are entire pieces of collective memory that remain excluded: undigitized archives, marginalized voices, minority languages, communities that lack the resources to document their own history. The shift to digital has profoundly transformed what is remembered and what is not, multiplying materials but also the risks of information saturation and "noise."
Artificial intelligence can help organize and make this sea of data navigable, but it can also alter the perception of reality if left unregulated. In this context, reflections like those on digital emotional anesthesia become crucial to understanding how the content that is resurfaced for us influences not only what we remember, but also what we believe we have experienced as a society.
Practical Examples: What's Happening Today?
To make this scenario concrete, we can look at three types of cases:
- Digitized Historical Archives and AI: Many museums, libraries, and research centers are using AI to index and make large digital collections searchable. This ranges from the documentary heritage supported by UNESCO to sophisticated open-access academic databases that use machine learning to connect new forms of historical testimony.
- Social Media as "Connected Memory": Social media create a form of "connective" memory, where personal memories intertwine with public ones through shares and hashtags. Algorithms decide what to resurface in "memories of today" and which narratives are reinforced over time.
- Misinformation and Recent History: False narratives about elections, pandemics, or conflicts can circulate and solidify rapidly as "public memory." As explored in our article on Fake News and AI: An Information War, Artificial Intelligence can be used to manipulate our perception of the present and, consequently, of the recent past.
Key Points
- Digital collective memory arises from the intertwining of institutional archives, private platforms, and recommendation algorithms that select what is visible and what remains hidden.
- Artificial intelligence can facilitate access to global documentary heritage, but it can also amplify biases, misinformation, and collective "false memories."
- There is not yet a truly neutral global archive: although international protection projects exist, many community memories remain on the margins of major digital systems.
- As citizens, we can develop awareness of algorithmic mechanisms, choose reliable sources, and support initiatives that preserve the plurality of memories.
FAQ
1. What does "algorithmic collective memory" mean? It means that a growing part of our shared memories – historical events, social narratives, symbolic images – is selected, ordered, and resurfaced by algorithms. It is no longer just schools or TV that decide what we remember, but digital platforms that optimize feeds and results based on behavioral models.
2. Is a global digital archive really possible? From a technical standpoint, it is possible to archive enormous quantities of documents and make them accessible anywhere. But a true global archive would require shared, inclusive, and transparent criteria on what to preserve, how to describe it, who governs it, and how to prevent exclusions and distortions.
3. How can AI distort collective memory? By amplifying distorted or false content, systematically excluding certain voices or events, and rewriting the past through syntheses that oversimplify historical complexity. These effects are particularly dangerous when they act invisibly within platforms we use every day.
4. What can we do as individuals to protect digital memory? We can cultivate "algorithmic literacy": understand that what we see online is filtered, verify sources, and support pluralistic projects. We can also choose platforms that offer more transparency, applying the principles we suggest in our guides on Digital Privacy in the Algorithmic Age.
5. Why do UNESCO and institutions talk so much about digital archives? Because documents, photos, and recordings risk being lost without long-term preservation programs. Initiatives like "Memory of the World" seek to ensure that at least part of this heritage remains accessible as a public good.
Conclusion: Towards Which Collective Memory Do We Want to Go?
The idea of a great global memory archive is fascinating: a digital place where humanity safeguards its traces, accessible to anyone, at any time. But if we look closer, we discover a fragmented landscape, governed by different interests and traversed by algorithms that filter, hierarchize, and reinterpret what is remembered.
As UNESCO writes in its manifesto, "the documentary heritage of the world belongs to all, should be fully preserved and protected and should be accessible to all without hindrance". A principle that today clashes with often opaque and privatized digital infrastructures.
The real question, then, is not just "will we have a global archive?" but "what kind of collective memory do we want to build with the help of algorithms?" Choosing means participating: informing ourselves, discussing, and supporting projects that treat memory not as a simple database to monetize, but as a common good to cultivate together.