Meta Morsels 2

A selection of new and notable writing on metascience

Following our inaugural post, this week’s selection is still dominated by AI since that seems to be commanding most people’s attention (even apart from all the latest doom and gloom about AI-mediated human extinction¹). But we have also found space for an ambitious proposal to facilitate the uptake of research on inequalities by social and economic policymakers.

Alondra Nelson: AI as Social Science Question, Object & Tool

Though it was quoted extensively in the Substack post from Henry Farrell included in last week’s Meta Morsels, Alondra Nelson’s illuminating essay is well worth reading in full. Nelson, who is the Harold F. Linder Chair and Professor in the School of Social Science at the Institute for Advanced Study and served as a science policy director in the Biden White House, builds a compelling case for how the social sciences should face the challenges and opportunities presented by AI. Her prescriptions apply equally well to the metasciences.

“AI models, tools, and systems pose three interrelated imperatives for social science: they demand renewed attention to social theories of how technology, human experience, and social order are entangled; they require study as objects of inquiry in their own right; and they offer capabilities that may transform—or upend—the practice of social investigation itself.”

Of particular note is her warning against reliance on closed proprietary AI models:

“There is a strong case that social science should refuse proprietary AI models entirely. The epistemic requirements of scholarly inquiry may be fundamentally at odds with profit-driven, black-boxed systems whose training data, architecture, and governance remain inaccessible to most researchers. Developing such alternatives requires investment and collaboration, but the payoff is research infrastructure that serves scholarship rather than market imperatives. This is not antitechnology sentiment but a defense of what distinguishes research from consultancy: the imperative that claims about the social world be contestable and answerable to communities of inquiry.”

Giorgio Gilestro: The Weimar of knowledge

Giorgio, an Associate Professor of Systems Neurobiology at Imperial College London (as well as being a friend and former colleague), also takes up the theme of openness in a trenchant and provocative blog post examining the implications for open science of the tsunami of AI-generated material that is overwhelming our systems of scholarly communication.

“Open science was, at bottom, a bet that if you removed the paywalls and the gatekeepers, the work would find its readers on its own merits. A reputation economy finally voids that bet. Free access is worthless when the binding constraint is not access but attention, and attention is allocated by brand. To be honest, we are already seeing that trend and AI will only accelerate. We will have open archives that nobody reads and closed reputational cartels that everybody does. That is a worse settlement than the one we are leaving, and it would be dishonest to present it as an equilibrium anyone should welcome.”

All may not yet be lost, however. In the age when text can be generated so easily, data becomes the most precious resource. To ensure that valuable data remain open, Giogio argues that the focus now should be on open source hardware:

“the real asset in the new regime is the instrument and the dataset it produces — and, crucially, the tacit knowledge of how to make the instrument produce something trustworthy. Whoever owns the apparatus owns the only input that is still scarce. This is why, in an age of technofeudalism, open source hardware development is an act of armed resistance.”

Stiglitz et al: Why we need an International Panel on Inequality

One of the main aims of publicly-funded research is to address societal challenges, and one of the greatest of these is inequality. As we have already explored on Go Meta!, the frictions of inequality create distorting power dynamics not just in society at large, but also the funding and agenda-setting of research. In a recent Nature Comment piece, Joseph Stiglitz and an international group of co-authors² bemoan the fact that research linking “higher income inequality to lower social mobility, poorer health outcomes and slower economic growth” has had relatively little impact on policymaking around the world. They propose the creation of an International Panel on Inequality (IPI) that would operate similarly to the International Panel on Climate Change. Their ambition for the IPI is to coordinate and standardise research on indicators of inequality and experiments in policymaking:

““Policy effectiveness depends on many factors, including state capacity, political coalitions, institutional arrangements and interactions with other policies. Policymakers need to know not just whether a measure is associated with lower inequality, but also how it worked, under which conditions, whether its effects endured and what other consequences followed.

An international assessment should create a systematic way to learn from both success and failure. Its aim would be to distinguish conclusions that are relevant to many countries from those that are specific to particular institutional and political contexts. It should also make evidence about implementation more accessible to governments, social partners, civil society and the wider public.”

In an accompanying editorial, Nature endorses the proposal but cautions that different perspectives will need to be included to counter the inevitable political headwinds. A companion Comment piece by Eoin McLaughlin argues that, to be properly sustainable, economic policies to address inequality must absorb the principles of intergenerational justice and be informed by methods that can properly track ‘inclusive wealth’, “a measure that focuses on three different forms of capital — produced (manufactured), natural (environmental) and human (skills and knowledge).”

1. For a dryly humorous but nonetheless insightful take on the brouhaha about the AI apocalypse, see Martin Robbins’ Substack post ‘The AI Apocalypse is Extremely Boring Now’.

2. Full author list: Joseph E. Stiglitz is a professor of economics at Columbia University, New York City, New York, USA. Adriana E. Abdenur is co-president of the Global Fund for a New Economy, Rio de Janeiro, Brazil. Winnie Byanyima is executive director of UNAIDS and an under-secretary-general at the United Nations, Geneva, Switzerland. Jayati Ghosh is a professor of economics at the University of Massachusetts at Amherst, Amherst, Massachusetts, USA. Imraan Valodia is a professor of economics and director of the Southern Centre for Inequality Studies at University of the Witwatersrand, Johannesburg, South Africa. Wanga Zembe-Mkabile is a professor of public health at the University of the Western Cape School of Public Health, Cape Town, South Africa.