The recent federal government announcement to secure long-term funding of Trove is not a win solely for the National Library and for Australians, but for the world, especially given recent developments in generative artificial intelligence, that is AI that generates content.
Subscribe now for unlimited access.
or signup to continue reading
In 2009, the Australian free digital online collection Trove was born.
A brilliantly conceived collaboration between libraries and archives across the country, Trove was improbably brought to life on a shoe string budget.
Almost 15 years on, Trove is not just a shiny, digital veneer on analogue stacks and repositories, but a prescient service, evidence of libraries continued leadership in information economies.
It is the past, present, and future of libraries in this moment when we face scaling of generative artificial intelligence services, a looming Dead Internet and an urgent need for societal retraining in information literacy.
Every day brings more news about generative AI services and how they appear to magically gather and synthesise information, with uneven and untrustworthy results.
Behind the deceptively simple user interfaces of ChatGPT, Google Bard, and similar services, Large Language Models are trained on vast amounts of digital and digitised information.
The relationships between the data on which they are trained and the outputs they produce are critical in current debates about the ends to which we use these services and the confidence we place in their outputs.
That is to say: how do we avoid - or at least know - when we are dealing with garbage in, garbage out? As the generative artificial intelligence services become bigger and more powerful, we stand on the brink of a new world.
As we face the unknown it is comforting that we are investing in archival institutions for our future.
As we face the unknown it is comforting that we are investing in archival institutions for our future.
The data created and managed by the robust network of libraries and archives that underpins services such as Trove provide not only the basis for generative AI experiences in service of public interests, but a future of responsible and trustworthy knowledge creation, including new models for our relationships with libraries as custodians of, and midwives, to knowledge.
Such investment is particularly critical when the internet no longer offers the quality of information we have come to expect.
Search results are already bloated with output from advertisers and content mills, employing search engine optimization and AI tools to manipulate outcomes.
With the rise of generative artificial intelligence, which can endlessly churn out derivative text and images, we already imagine a future where the web becomes a blurry facsimile of what it once was.
The Dead Internet Theory, which claims the vast majority of content on the web is created by and for AI bots, could shift from the domain of conspiracy theorists to that of information literacy trainers in our educational and cultural institutions.
The library is ideally equipped and placed to steer us through a number of near and far future challenges and opportunities for shaping responsible and trustworthy information economies in the coming century.
With indications of the coming flood of generative AI experiments built on historic data sets, the track record of libraries as sites of innovation and new information provides confidence in the role they must play in managing such a significant change to our information landscapes.
Libraries were a key player in the information management revolution of the 20th century.
From microfilm and electric checkout to pneumatic tube systems and computerized catalogues, libraries have long been early champions of new services based on emerging technology capabilities.
So it makes sense they play a key role in the 21st century.
- Alexandra Zafiroglu is a professor of cybernetics at the Australian National University school of cybernetics and a fellow of the ANU Futures scheme.
