Beyond Theory - Interview of Randi Cecchine

The Beyond Theory interview series is a recent project by the PAAG-ICA group designed to offer photographic and audio-visual management-related content, enabling operational possibilities through a pragmatic approach. The main goal is to interview relevant and highly experienced professionals, who deal with different aspects of the audio-visual and photographic workflow. This is the second interview of the series and it was an asynchronous email interview. 

Interview of Randi Cecchine. By Juan Alonso

Randi Cecchine has a background in documentary filmmaking, media education, and educational film distribution. She is passionate about the potential of Artificial Intelligence in A/V archives. In the context of a Master’s programme in Preservation and Presentation of the Moving Image at the University of Amsterdam, her academic interest focussed on how institutions are implementing the use of AI systems. She was also intern at the Netherlands Institute for Sound & Vision. As part of her academic research, she recently wrote an interesting and highly recommended series of blog posts offering examples of how AI is being used for search and discovery, creative reuse, education, audience engagement, and collections as data. During the interview, Randi addresses technological innovations in the management of audiovisual archives very attentively, while taking into account profound ethical, philosophical, environmental and human rights issues.

You have got a very interesting hybrid profile, considering your academic background and extensive professional experience both as a video-maker, and as a lecturer in visual and media studies. Simultaneously, you are studying for a Master’s in Preservation and Presentation of the Moving Image. Do you consider your approach – more pragmatic or more re-use focused – different from that of an archivist? In what would you differ?

Thanks so much, Juan, for the opportunity to be interviewed and thanks for this rich question - it has given me a chance to reflect upon my background and interests. 

My experience as a filmmaker, freelance video producer making videos for organizations, an educator, and an acquisitions manager in an educational film distribution company is what led me to my interest in archives and AI. My thinking is grounded in a deep engagement with making and teaching, and also by all the wildly diverse topics I’ve learned about in the course of my career. For example, I’m very influenced by what I learned working on a project of the Center for Children and Technology called Math For All, a program created to help teachers teach math to kids with disabilities through a neurodevelopmental framework. I love to learn and I love to learn about how we learn.  Media and archives are part of our literacy and cultural heritage.

When I was working as an acquisitions manager for a streaming educational distribution platform I began learning about metadata and databases, and it was then that I understood that the work we were doing translated into the work of archives. In the course of the acquisitions pipeline, from selection to ingest I found myself often working with large spreadsheets of videos and thinking about how I was learning from the metadata about a collection, even if I hadn’t personally seen the videos. I wondered how the aggregated information of collections could be used as data, and I also wondered how machines might help us process this information.  

I was motivated to enter the University of Amsterdam’s Preservation and Presentation of the Moving Image program to help me understand the theoretical background of archival work and with a very open approach to learning about AI and archives. I took my own curiosity as a starting point - just as I often have with making documentaries. When I applied to the program I didn’t know that the Netherlands was a leader in developments around AI and archives.  I was delighted to meet some amazing scholars and practitioners here who have helped teach and guide me. They, and the people that I interviewed in the process of writing my blog series, helped to inform my approach.

I want to return to your question about if I consider my approach more pragmatic or re-use focused and how it might be different from an archivist. I think it is important to say that I’m a professional making a mid-life career transition and a transition from New York to the Netherlands. I am a master’s student and the blog series was made in the context of an internship. It was the first time I’ve ever done a research project with a written output rather than a video.  I say this not to downplay my research - I think it is a quite comprehensive view on the state of the field - but to say that it was driven by a sort of “beginner’s mind” approach and might not fit into any existing categories. I’m very curious about the cultural differences between archives - especially between the US and Europe, and how these differences influence decisions made in archives about what types of tools to implement. I see very different approaches to the use of AI in archives, and those approaches are grounded in pre-existing values in archives.     

So I think that my hybrid approach, and especially having spent tremendous amounts of time shooting and editing video, gives me an understanding of the many, many layers of information that are housed in AV collections, and a deep awareness that all creation is always biased and grounded in individual style, culture, and belief. I think that this understanding - something you understand deeply from editing documentary materials, is essential in understanding the complex relationship between machine learning and training data, and the human decisions that get made in the process of implementing AI into archives.

I’d like to go back to your question. You asked “Do you consider your approach – more pragmatic or more re-use focused – different from that of an archivist? In what would you differ?”

I think I’ve tried to approach this research from the point of view of the archive, but my own perspective is probably more broad than that, since I am not an archivist. I’m a generalist as a thinker, and always interested in multi-disciplinary thinking. I’m interested in how projects arise, what values are expressed in the choices made, and how we can become more aware of these factors.

In relation to the previous question, do you think that an audio-visual archivist should have technical knowledge of production and editing? Why?

I don’t think I’m in the position to say what audio-visual archivists should know but I certainly think it helps if archivists have some experience with production and editing on both a technical and also creative level. I think there are some conceptual aspects to production that may only be understandable when you’ve done it yourself.

For example, one of the things I’ve been thinking a lot about is how archives handle the source materials created in the production of a documentary. A lot of documentary filmmakers amass large troves of content for their films and that content may be very interesting for users. I have a 50 minute documentary with 350 hours of source material. I think that these original materials such as media files, film or audio assets, even still photos - have value in and of themselves - and that the original edit project files are already organizing the materials. I know that archives can’t take everything - but I think we will see more work being done to establish conventions for how to save the media files and the edit decision lists (project files) in some form so that future users can access not only the media but the project files. I think this is especially important for TV archives and also for organizations such as non profits doing their own production work.

When I’ve tried to discuss this with some people in archives I’ve found that they think about the source material as “outtakes” - like you would with a fiction film. Like they are the “bad” parts - the scraps. I don’t see the source material for a documentary that way at all.  When you make a film, you have the luxury of learning from all the content and all that learning informs the final product. The final product - what you are able to include - is always a very slim representation of the reality of the topic - filtered through the makers’ own perspectives and whatever format dictates they may have. For people who are interested in the topic, the source materials may be of much greater value than the final product!  I documented many years of the anti-war movement of the mid 2000’s in the US and I think that footage has its own value. As I was struggling to edit the film I found some kind of comfort in thinking that it might one day be valuable in a future archive.

I think it would be helpful if there was more dialog between archivists and makers about their processes. I was happy to edit a series of videos called Meet the Archive for the Eye Filmmuseum’s Collection Centre about the work of their curators, many who work closely with filmmakers. I think it is really important for archivists to be in some way connected to the process of making work, because it increases the respect for the collections and their potential. I have a strong feeling of camaraderie especially with independent documentary filmmakers, and I think their input could be incredibly valuable in archives. In the best case scenario, archivists are loyal caretakers to the sacred work of the filmmakers.  

We know that so-called film-researchers work in audio-visual productions to search for archival images of historical nature. What happens to the raw footage of an audio-visual production of a digital film? Who is taking care of it, and how is it managed? Are logistic profiles those in charge of downloading, verifying, and documenting the images? What happens to the digital-born raw files when the movie comes to an end?

As I mentioned above, this is a topic that I’m extremely interested in, and I wrote a paper exploring it. I also had the pleasure of working on a case study at the Eye Filmmuseum’s Collection Centre on the source material of documentary filmmaker Frank Scheffer. I don’t think there’s any single answer to your question. Each network or production company handles this differently and I don’t think there is a clear position about this in any of the archive organizations represented by the Coordinating Council of Audiovisual Archives Associations CCAAA. (Readers, please correct me if I’m wrong!) I think that besides the technical challenges that saving these materials in an organized fashion presents, there are copyright restrictions on how the materials can be used. It is a pretty big can of worms and something I’d love to see talked about in much more depth.

In historical archives and museums, one can find home movies made by amateurs documenting the daily life of people, usually with high purchasing power, from the past. Nowadays, the audio-visual recording of everyday life has been democratised, and basically everyone has got a device to capture personal moments. In addition, the way of recording seems not meant to preserve the personal or family memories (to which we do not usually resort/revisit as before), but seems more intended to share experiences immediately on the social media platforms, such as Instagram or WhatsApp. This makes the capturing to have a more ephemeral component. Perhaps this is a too theoretical or futuristic question, but what do you think will happen with all this born-digital audio-visual files? Do you think they will ever reach the historical archives? Will back-up files of conversations, images, and videos from social networks, such as WhatsApp or Telegram, be received? What should be the role of the archivist in this regard? Should archivists care more about this?

I think this is really an essential question for our times and I love this kind of future-oriented question. It reminds me of the Futures Literacy Framework from UNESCO. For me, the question puts into focus the role of the big tech firms in managing our personal data/archives for the long term. If we think about the photos and videos on our phone, currently most people are using some sort of cloud service like Apple to save them, and may be uploading some to platforms owned by Meta or Google. I think the way that we use cloud services and outsource our own archives to these companies represents the fundamentally flawed relationship that we are in with these companies. We allow them to use the data to train algorithms, we lose track of it, we become very passive and they take whatever they want. We trust them when we really have no reason to. We also build up huge server farms that guzzle energy to keep alive these memories that we might never revisit. I don’t see how this can be sustainable into the future in which energy will be much more costly and environmentally damaging. Do we really need to keep digging for ancient fuel to keep our digital trash?

I recently read a very disturbing and insightful article called This Private Equity Firm Is Amassing Companies That Collect Data on America’s Children. This quote struck me as an essential way to think about the value of our personal data: "You don’t pay the oil company to come pump oil off your land; it’s the other way around." Alex Bowers, Teachers College, Columbia University.

I know that I’m drifting from your question, but I think it is really essential to remember that the reason we share our personal information on social media is entirely disconnected from the business model that tech firms have with our data. They want it all, and they want to be able to do anything they want with it.  So the saving of it, and who has rights to analyze it, to use it to train algorithms, to build algorithms with intentions that we couldn’t even imagine - this is the value of that data.  We have been willingly giving it away, sometimes worried about personal privacy - which is not my concern. My concern is that we are enriching companies without understanding what they are doing with our data. We are behaving in a very naive way.  I also personally lost a YouTube channel where I had been saving a lot of my video work. There was no way to get it back from the “squatters” who took it over.  I am embarrassed that I trusted this company to hold my archive. Maybe someone reading this has a kind human connection at Google who wants to recover the account for me?

But to return to your question - I think these are explorations that archives are in a very good position to lead. I know that some archives are archiving websites, for example of broadcasters.  

My dream would be that some people would start to build tools that “normal” - non-technical people could use to rescue their archives from big tech and organize them locally with the help of some AI. I liked an idea that Anne Gant from the Eye Filmmuseum recently shared during an AMIA workshop on archiving for filmmakers. She suggested that we give one another hard drives as gifts! 

Sometimes, when I talk to young people about audio-visual archiving, they ask questions about analogue material that make me think they have poor technical knowledge about old films and magnetic tapes. This is critical, considering the large amount of audio-visual heritage on magnetic and film media that must constantly be transferred onto digital media. Do you think there is a deficit in the transfer of generational knowledge?

First, for anyone interested in this question in relation to magnetic tape, I would direct them towards the IASA/UNESCO Magnetic Tape Alert Project.  

Yes, I think this problem of generational knowledge transfer is very significant and I’ve heard about it from various people, including companies offering digitization services.  Aside from the knowledge of what the formats are and how to use the equipment, I think there’s a larger problem in terms of the expertise needed to maintain the equipment. I think we’ll see a heavy drop in the ability to digitize magnetic tape over the next 20 years as the equipment becomes less available and the expertise to maintain it becomes more scarce. One of the fun and positive things though is that 3D printing makes it possible to get/make parts that would have been impossible before.  

I’m delighted these days by how young people are interested in analog tech such as vinyl and film cameras. I've never heard a young person excited about shooting Hi-8 or C-VHS though! Sometimes I can’t believe how many years we shot on these very poor quality carriers. 

I would like to think that those of us who experienced/worked with the transitions between formats would have some expertise to share that would be valuable.  

Ah, this reminds me of an amazing project called ADAPT: Researching the history of television production technology. I was involved with organizing and describing the videos to be placed in the Europeana archive. The project brought together veteran BBC crew to reenact the work they used to do in the 70’s, 80’s and 90’s using the original video and film cameras and editing equipment. The project documents the work and the crew’s reflections.  I learned so much about the technology watching the videos, and I especially learned how much expertise early field (outside broadcast) cameras needed to work. I really recommend this series for anyone who is interested in the evolution of technology. I like ideas around Media Archaeology. I think there’s so much delight in having a closer relationship to the working of equipment.

Recently I was in a webinar with a young student who was complaining about the amount of processing time it took to colorize footage. I responded with such an old person’s story about how long it used to take to process a roll of 35mm film - rolling it on a spool in the dark, all those chemicals, and then the time to dry it. One had to practice a lot of patience and I think that processing time was important to the work itself. Our expectation of immediacy is a very new phenomenon.  

In one of your blog stories, you talk about a change in the role of audio-visual archives, of a kind of transition from an archive as a space for information custody towards something different, more related to personalised diffusion. Could you share more thoughts on this, and will those be the archives of the future?

This is a tricky question. I think I was talking about a project where an American public broadcast archive was looking to use AI eventually to help personalize content recommendations. I followed up with a quote by a European public broadcaster talking about how public broadcasters actually have a mandate to broaden people’s tastes rather than to funnel them into “filter bubbles”. I really like this concept of “taste broadening” that I’ve heard of in Europe.

I’m someone who generally doesn’t follow mass-tastes, and I’m also someone who doesn’t have a lot of trust in media companies.  

I find this push towards predictive algorithms very troubling. AI replicates the bias that the individuals who create it and the data it is trained on contains. We lose our autonomy when we don’t realize that the information being shown to us is being filtered through what the algorithms have decided we are. As I write this I picture some of my early memories of being in a library exploring both the card catalog and the shelves. Imagine if I was only shown what the librarian thought I would be interested in!  I would have never stumbled upon books that have had a deep impact on me.

The risk is that we become the product that the algorithm predicted us to become. This is really the philosophical question of our times and I’d love to hear of anyone writing or thinking about this.  I always loved the book “Four Arguments for the Elimination of Television”. I think we should play around with similar ideas in today’s world.  

When I think about a product like Spotify I wish that the wisdom of the people who created the algorithms could be shared back to me as wisdom rather than prediction. I wish that I could get Spotify to help me understand the elements of music better, to even understand why I like what I like, and to recommend some music that I’ve never heard of. I really like the idea of Hybrid Intelligence that some Dutch researchers are exploring. Thinking about how Artificial Intelligence and Human Intelligence can work together placing humans in the center is really exciting to me.  But at the end of the day for me the question always comes down to the values and agenda of those making the products. I think that the potential for AI becomes very narrow when it has to align with corporate interests. It is why I think the landscape of audiovisual archives and AI is so fertile - because archives have training data and can think about their role to serve the public without being driven by market demands.

The daily production of data is enormous, and when we talk about audio-visual documentation, we are moving into other dimensions. It is said that we “live beyond our means”, in the sense that we produce beyond our storage capacity due to the high costs associated with hard drive technology and LTO tape. How do you think the digital storage sector will evolve? Do you think a new technology will develop? What will lower the costs?

I’m really not an expert in this area but it looks like cloud storage solutions are becoming extremely affordable and are offering many different tiers that allow for further decreasing costs, for example for items that won’t be accessed often. I know that a lot of archives in the US are already using commercial cloud storage - including the US government. I’ve been curious about how local regulations in the EU have slowed this down. I believe that many public institutions in the EU have regulations that their archives need to be stored in their country. A few years ago there were no Amazon servers in the Netherlands, for example, but now there are. A lot of the local infrastructure was created in that gap of time, and storage is handled locally. I am curious to see if Europeans will eventually move over to commercial cloud storage providers because of the competitive cost. But to me the big question is what does it mean to lose local control of your collections? I can’t answer that question but I think it is an important one to explore as we think about the long term futures that we are building towards. What is an archive if it actually doesn’t hold its own archive?

Regarding the future: I have learned and written a bit about new storage technologies like Microsoft Research’s Project Silica or storing files in synthetic DNA, and I’m fascinated by these ideas. I attended an interesting webinar in 2020 called Digital Storage Futures where there was a presentation about a decentralized storage network called Filecoin that works in some ways like the blockchain. I liked the concept.

I’m sure there will be many incremental advances and I think all sectors need to quickly pivot our thinking about digital storage from thinking about access and money to thinking about  energy consumption and the environment.  I’ve been participating in the Europeana Climate Action Steering Group to explore these issues.  I’m often irritated by the language of “the cloud” which can make us forget that “cloud” storage happens in multiple physical locations consuming large amounts of resources. One of the most provocative things I’ve read on this topic was Streaming video, a link between pandemic and climate crisis (Journal of Visual Culture & HaFI, 2) that broke down the carbon footprint of  streaming the Netflix Series Tiger King over a specific period of time and found it to be the same as the electrical consumption of Rwanda in 2016. I think that this type of analysis really helps to make tangible that our digital activities have an impact. As we move into a future of unknown energy availability, we really need to get real about this. I like to say that there are two kinds of climate change denialists - the people who deny the existence of climate change and the rest of us. Denial is a natural human response to a situation that is beyond our ability to imagine. We need to get real.

Nowadays there is a lot of talk about Artificial Intelligence and Archives. You, in addition to being passionate about AI, are currently researching how AV archives are implementing the use of AI systems. I have the impression that sometimes we confuse process automation with artificial intelligence. How do you define it? Are there different types of AI? To put it simply, could we say that process automation is a robot that does things, and the AI is a robot that does things and also learns and takes decisions?

This is a good question that makes me reflect on how we use the term AI. I think that each person has a different concept of what we think AI means. I had the pleasure of attending the ACM Multimedia Conference in Nice in 2019 and I think that being around hundreds of computer scientists who were presenting their projects helped to ground my understanding of Artificial Intelligence, or Machine Learning, as specific tools created by specific people with specific goals. The algorithms can do some “learning” that they are programmed to do and then these algorithms can be used together to produce a complex result.  I think your distinction about process automation and AI is probably accurate, although I could imagine a system which “learns” from the automation. These are the places where people become empowered to figure out what AI can do for them and become more involved in the design of systems that they actually need rather than waiting for the developers to come up with it.

In general, I think we all need more “AI literacy” to counter some of the narratives that would make us believe that the threat of AI is its sentience. My cynical side actually thinks that the media companies that are feasting on our data to create AI enjoys feeding us fearful narratives about sentient AI just to put us off their trail. 

In the last FIAT/IFTA World Conferences, many examples of AI and feature extraction in TV archives have been shared. In general, it seems very widespread, and during your research, you have interviewed very relevant people in the field of audio-visual archives who work in large archives. However, in your opinion, what is the real degree of actual implementation of AI in TV archives? Do you think AI projects in the archives are more exceptions or norms? Why?

I am more familiar with archives represented by AMIA and IASA than the TV archives in FIAT/IFTA, but my sense is that there’s a lot of work being done in TV archives. One of the things I had to learn when I first started looking at conference presentations about AI projects in archives is that a lot of them were actually prototypes and not fully operational. I’m almost embarrassed by how long it took me to figure that out! There’s a lot of hype around AI and I don’t think people are doing a very good job of giving a clear picture of what is really going on.  

My sense is that the actual state of the field is filled with a very diverse set of approaches - with different types of archives implementing wildly different types of tools and systems. The tools that come with MAM or DAM systems I assume are improving, and then it’s just a cost question for what can be implemented. I think the EBU DATA groups are probably the best place for this kind of inquiry.

In audiovisual archives, we have digitised (and keep digitising) massively, we have created long-term digital preservation systems and now we are already dealing (apparently) with the management of born-digital videos. It seems that, in general, we are in a new phase marked not so much by preservation, but by access and re-use. We see more and more productions brilliantly using stock footage to not only document themselves, but even to combine archival footage with new footage, such as in the movie “Munich” or in recent and successful series, such as “Wild Wild Country” or “The Last Dance”. However, beyond re-use in new productions, AI opens up new possibilities. Which are the most relevant ones? Do you think archives are a good space to implement AI? Why? Could you share interesting case studies in this context?

My blogpost The Potential for AI in Audiovisual Archives explores various uses for AI in archives and looks at the idea of Collections as Data. As a filmmaker, I am always thinking about the learning that happens through the filmmaking process as I engage with transcripts and I review content over and over and over. I like to think about what knowledge or stories are in our archives that can be revealed as we are able to analyze content at scale.  Speech-to-text (or automatic transcription)  is the simplest tool to explain, and the potential for it is immense. If an archive has access to transcripts of its content, and I think this is especially relevant for documentary, news, information, educational content - then those transcripts can be searched, and new kinds of questions can evolve.  I think a nice example of a tool offering access to new types of information is the speech segmentation tool created by INA. The tool has a lot of functions because it helps separate different types of audio that can be analyzed by other tools, and it also can analyze the gender of speakers in an audio recording. This opens up new potential to look at our television history and understand how it has changed over time.  

I also really like the work being done by the Netherlands Institute for Sound and Vision and the CLARIAH media suite working with archive users to create Data Stories.

Does AI make sense in all the archives? In other words, there are small municipal archives managed by two or three people who have few financial resources, but who are aware of the importance of AI and do not want to leave behind these technological advancements. What advice would you give them? Does low-cost AI exist?

I think first of all it is important to recognize that there are many archives in the world that struggle to find the resources to manage their basic functions.  Implementing AI tools is absolutely not a necessity and if archives are struggling with preserving and digitizing backlogs of film or magnetic tape, they shouldn’t think they need to get ahead of AI.  AI will keep developing and will be more accessible in the future. I think the wealthier archives are in a position to do the research to implement some great ideas that will help smaller and struggling archives to figure out what is important to implement in the future.  
I’ve been talking with an archivist in Tunisia who has been testing speech-to-text tools and he found that while many tools can do a good job with Arabic, there aren’t many doing a good job with the Tunisian dialect of Arabic. I think that if there are opportunities for funding, archives (with their data sets) are in a good position to be able to partner with academics and private industry to create AI tools.

There are a lot of open source - “free” AI tools but they require staff who know how to implement them, and enough computing power to run them. I think these questions all come down to what tools would be most helpful, and what are the resources required to utilize them, and what kinds of funding options are there. 

Let us imagine a medium-sized audiovisual archive that wants to start implementing Artificial Intelligence projects. Are there protocols, guides or standards for this? What aspects should be taken into account in the selection of requirements? What tools would you recommend?

One of my favorite interviews that I conducted about implementation of AI was with Meemo, the Flemish Institute for Archives. There were 4 people in the call, and each of them brought a different set of insights and expertise to the table. At that time they had not yet begun to implement any AI tools, but they were deeply knowledgeable about the myriad of technical and ethical and rights questions to be addressed in order to do so.  I really liked the way they worked and thought together, and how they looked towards the experience of other archives to help their journey. They made me realize that the “early adopters” actually are at a disadvantage - they may receive the benefits and prestige of using a new technology, but they also have to make all the mistakes and spend all the money! The “later adopters” have the benefit of all those experiences plus the exponential advances in technology to help them.  Some countries, like the Netherlands, have made the development of AI a national priority with organizations like the NL AI Coalition, and I think the lessons learned here can be a huge benefit to people in other parts of the world.

As far as I know, there are no particular protocols, guidelines or standards in place through the archive organizations that I’m familiar with, although EBU may have some specific ones for broadcasters.  

I would recommend that any archive interested in exploring AI projects reads the comprehensive report commissioned by the US Library of Congress Machine Learning + Libraries A Report on the State of the Field and assemble an internal group of people with various kinds of expertise to discuss.  I think it is vitally important that conversations happen between “tech” people and “non-tech” people, and that there is an emphasis placed on developing a shared vocabulary that encourages the participation of “non-tech” people in the conversations and decisions about implementing AI.

I would suggest that archives identify tools that would really benefit their unique collections and user needs, and figure out how their systems are setup to either be able to run these tools themselves, select them as additional features in existing DAM/MAM systems, or work with an external vendor such as Grey Meta to analyze material. AVP in New York has a webinar online Finding the Right AI Tools for DAM: AVP’s Human-Centered Evaluation Framework that explains the work they are doing helping organizations in this process. There are a lot of people who I respect who push for only using open source tools, but I think sometimes they forget that these tools require a lot of knowledge and resources to run.  I think the decision to do this work in-house or to work with external companies is an important topic to explore in depth.  It has to do with what kinds of resources archives want to invest in and what types of human talent they want to bring on board.

Finally, what would you say to an archivist who claims that a machine can never replace human work in a digital audiovisual archive?

I agree completely! I think we live in a time where we have a lot of confusion about the value and purpose of human labor. Earlier I mentioned the series called Meet the Archive that I edited for the Eye Filmmuseum. I did this after I wrote the blog series on AI and Archives and  I found it thrilling to hear the curators speak about the knowledge that they have built about the collections and filmmakers over their long careers. I felt like I could understand the value of  the knowledge that these professionals have built over their careers, and the generous way they wanted to share the knowledge. It is so ephemeral - all this information in a person’s head!  And isn’t it the purpose of life?  

To me, the promise of AI is when it can help us to learn and build our wisdom and creativity even deeper and broader. For example, I am really excited to see how AI tools will be implemented into video editing software in the future and how this can help me organize materials better - but I have no interest in giving up on the joy of editing and getting to know footage deeply because it is actually the process of doing the work that I love!  I think we need to learn how to value human process and learning in order to figure out how to build tools that will complement us, not attempt to replace us.

But machines are already replacing people in many institutions. This changes the career trajectory inside archives, and I think we all need to be open and honest about this transition and find more spaces for communication.