Close
A new research area under the Future Heritage theme at the UCL Institute for Sustainable Heritage.
Background
Artificial intelligence will form the industrial heritage of the future. Through policy efforts such as the UK's Industrial Strategy, the algorithmic systems that comprise 'AI' will become a central part of our economy, the equivalent of the mills of the first industrial revolution, and just as much part of our national heritage.
These systems mediate information, they regulate choice, and they are changing the ways in which products and services are designed and delivered: as people become more aware of how AI systems shape the world, they become meaningful beyond the technical fields in which they originated.
Understanding these systems' part in shaping the present will help future generations better understand their world. Yet there is currently no discussion, in the UK or globally, of how these systems might be recognised as future heritage, and it is far from clear how such a heritage would be managed, from a technical or policy perspective.
Addressing this challenge to heritage policy presents difficult practical questions. For example: if 'artificial intelligence' is recognised as heritage, what are the boundaries of the sociotechnical systems that comprise it? Where are they located - are such systems national or supra-national? If protecting heritage means understanding the ways in which it changes and decays over time, what kind of heritage science is necessary for understanding how AI systems change over time?
If heritage policy is already concerned with intangible cultural heritage like languages or music, what lessons can be drawn from that experience? And if heritage must be protected, and must be accessible to the society that values it, how best can this responsibility be shared between public bodies and the companies and institutions developing and maintaining these systems?
In exploring these questions, and considering the many technological and conceptual challenges they raise, our understanding of what constitutes heritage and the nature of our response to new forms of heritage will be enlarged. But answers to questions like these are also relevant beyond the world of heritage policy.
Thinking of AI systems as future heritage offers a new way of thinking about their governance. It frames them as the creation of humans, positioning them as systems within our control. It emphasises our stewardship of them, highlighting our responsibility for their actions and effects. It draws attention to the persistence of these systems over time, raising questions about how they might be retired from service, or about the nature of their legacy (what has been the impact of the original PageRank algorithm, for example?).
Perhaps most usefully, thinking of algorithmic systems as future heritage might lower the temperature of debates about their governance, helping us see that AI is not something to fear but to celebrate, and that taking the steps necessary to make it accessible to future generations will ensure that future generations also recognise this value.
Research
The Algorithmic Heritage research area explores these questions through three key activities:听
- developing a theoretical and conceptual framework through drawing on relevant research in computer science, science and technology studies, heritage studies and other relevant disciplines
- undertaking ethnographic research to understand what thinking of algorithms as heritage might entail for firms and organisations creating and using them in their work
- working with policy and technology experts to understand how these lessons from the field might be translated into practical approaches for managing this future heritage
To learn more, contact Professor听Richard Sandford: r.sandford@ucl.ac.uk
