Charleston 2026: Securing a fair return in the AI economy

Libraries and publishers must recognise what they have built together – and what they stand to lose, says Rosalyn Metz
Libraries and publishers risk repeating the mistakes of the digital transition if they exchange the enduring value of their collections for one-off payments from artificial intelligence companies, Rosalyn Metz warned delegates in her keynote speech at Charleston Conference.
Metz, Chief Technology Officer at Emory University Libraries, urged libraries and publishers to recognise their shared interests and negotiate agreements that protect the long-term value of scholarly knowledge.
Her message was that organisations responsible for creating, organising and preserving knowledge must build reciprocity into their agreements with businesses seeking to exploit it. Immediate financial returns should not obscure the decades of commercial value that collections could generate – or the investment needed to sustain them.
“The time to negotiate reciprocity is not 20 years from now,” she said.
For Metz, this requires a reassessment of the relationship between libraries and publishers. Long accustomed to negotiating from opposite sides of the table, both now face technology companies with the resources to reshape the information ecosystem around their own interests.
Their response, she argued, must begin with recognising what they have built together – and what they stand to lose.
A shared history
Metz traced the relationship between publishing, libraries and public access back to the early United States, when circulating information was understood as essential to democratic participation.
Early publishers faced shortages of paper, unreliable communications and political pressure. The Postal Service Act of 1792, she explained, helped address these barriers by allowing newspapers to travel at heavily subsidised rates. Information distribution became a matter of collective investment. Libraries subsequently expanded access beyond those able to afford private subscriptions. Public funding and lending arrangements helped establish an infrastructure through which knowledge could circulate more widely.
Physical ownership was central to that system. Metz described how the first sale doctrine provided a foundation for libraries to lend lawfully acquired books without seeking publishers’ permission for each transaction.
Print also imposed practical limits. A particular copy could generally be read by one person at a time, deteriorated through use and had to be transported between locations. Those constraints helped support a balance between lending and commercial sales. Libraries widened access while publishers continued to generate revenue.
“For generations, library lending and commercial publishing really relied on that delicate balance,” Metz said, adding that the internet disrupted the conditions on which that relationship depended.
Abundant content, finite infrastructure
To explain the consequences, Metz referred to two economic characteristics: excludability, or the ability to restrict access to a resource, and rivalry, where one person’s use reduces what remains available to others.
A physical book can be restricted through ownership and price, while its use is constrained by the availability of individual copies. Digital information behaves differently: a file can be copied and used by many people without exhausting its contents.
However, the systems delivering that information have finite capacity. Servers, bandwidth, software and computing resources all require investment, and heavy use can affect availability.
This distinction has become particularly pressing as automated systems harvest digital collections. Even where information is openly available, the cost of serving it does not disappear.
“If you don’t scale your infrastructure or block the bots, users won’t be able to access the platform,” Metz said.
Expanding capacity alone was insufficient, she suggested. A colleague had pointed out that additional infrastructure could simply enable bots to consume more, driving costs higher again.
The result is a sustainability problem for institutions committed to access. Making a collection available can expose its host to demands that undermine its ability to serve researchers, students and other users. This demonstrates why assumptions inherited from print cannot adequately govern digital stewardship, she argued.
Governing the knowledge commons
Metz drew on the work of the political scientist and political economist Elinor Ostrom to explore how communities manage shared resources.
Successful arrangements, she explained, depend on more than access rules. They require a shared understanding of the resource, trust between participants, a long-term perspective and the ability to establish workable governance. Libraries and publishers historically possessed many of those characteristics. They understood physical books, maintained continuing relationships and operated within a framework that allowed each to make decisions about its own activities.
Digital distribution weakened those boundaries. The community responsible for sustaining information now operates within a much larger environment, populated by actors that may have little interest in its continued health. “How do we govern when it’s unbounded and it’s a digital ecosystem that includes external actors we don’t have control over?” Metz asked.
She argued that libraries and publishers had each contributed to the imbalance. Publishers had tended towards excessive enclosure, while libraries had pursued openness without adequately managing its financial and operational implications. Meanwhile, technology companies had occupied the space between them.
The challenge was therefore broader than resolving familiar disagreements about access or pricing. It concerned who would determine the conditions under which knowledge circulates, and who would benefit financially from its use.
The retreat into trusted networks
Metz placed these changes within a wider transformation of online life.
The early web, she recalled, appeared to offer an open, collaborative environment built around shared knowledge and community participation. Commercialisation, tracking and algorithmic optimisation subsequently changed that experience.
Drawing on discussions of the internet’s “dark forest” and the “cozy web”, she described how people had increasingly moved towards smaller, more trusted spaces: private messaging groups, newsletters, email and professional communities. In these environments, ideas circulate through personal recommendations and established relationships, rather than relying entirely on public discovery mechanisms.
Metz recognised this pattern in her own experience of publishing on Substack, where conversations with colleagues developed into writing that faculty members subsequently recommended to their readers.
Her wider point was that human communication and automated consumption were moving in different directions. People sought trusted networks while bots placed growing demands on publicly accessible resources.
That divergence further complicated the assumption that simply putting material online would produce a healthy information commons.
Contracts that shape markets
For a possible response, Metz turned to economist Mariana Mazzucato’s work on shaping markets around the common good.
As Metz presented it, markets are influenced by choices about procurement, intellectual property, investment and institutional design. Organisations therefore have opportunities to influence commercial behaviour through the conditions they establish.
She outlined principles including a fair distribution of risks and rewards, transparent accountability, shared goals, participation and collective learning.
Applied to scholarly communication, these principles invite questions about who pays for knowledge production and preservation, who captures the resulting value, and what obligations accompany commercial access.
Metz used the Oxford-AstraZeneca vaccine partnership as an illustration of how agreements could connect academic research with public objectives. In her account, contractual conditions around technology transfer and pricing demonstrated that universities could influence how private companies commercialised research.
The relevance for libraries and publishers lay in treating contracts as instruments for advancing institutional missions.
Organisations could negotiate for continuing returns, cost recovery or public benefits, rather than regard the immediate exchange of money and content as the whole transaction.
Avoiding the next dependency
Metz contrasted publishers’ development into analytics and data businesses with libraries’ continuing tendency to treat technology as a supporting function: “We, however, in libraries have continued to treat technology as back office.”
Some publishers had recognised digital knowledge as central to their business models, embedding services in research workflows and articulating their value through data and utility. Libraries needed to examine what their own technological infrastructure meant for their future role.
That reassessment was urgent as AI companies harvested research, cultural heritage materials, curated data and metadata.
Metz invoked Cliff Lynch’s criticism of the dependency created by licensing content rather than owning it, then identified a potential equivalent for artificial intelligence: “I don’t think we can afford the original sin of the AI era,” she said. “That sin will be signing away our collections for one-time payments.”
Companies could derive value from those materials for decades, she argued. A single payment would fail to recognise that continuing benefit or support the institutions responsible for maintaining the underlying resources. Libraries and publishers therefore needed to look beyond immediate budget pressures, sales targets and procurement decisions when negotiating AI agreements.
Sustaining the source of value
Metz also directed a warning towards vendors: their businesses depend on healthy academic institutions.
Research communities produce the scholarship and data that commercial services organise, analyse and sell. Undermining those communities would ultimately weaken the source of commercial value. “If academia is hollowed out, the well will run dry,” she said.
For librarians, the corresponding responsibility was to reconcile openness with active stewardship. Protecting infrastructure and negotiating sustainable terms could support the mission of access by ensuring collections remained available over time: “Making something open doesn’t mean leaving it unprotected.”
The questions extend to creators and donors, whose work institutions hold in trust, and to members of the public who could otherwise find themselves priced out of tools built from collective knowledge.
Metz’ closing appeal was for libraries and publishers to negotiate from a clearer understanding of their interdependence and their continuing contribution. “Let’s build contracts that honour the true ongoing value of what we are in the business of helping to create, preserve, and steward,” she concluded.
