2024 BCCP Conference Review

Focusing on the prospects, challenges and regulation of AI, the fifth annual Conference and Policy Forum of the Berlin Centre for Consumer Policies (BCCP) was held in Berlin on June 14th, 2024.
Touching upon an issue at the forefront of current European and global policy debates, more than 100 participants, including academics from economics, law, computer science, and ethics, policy makers, professionals, BCCP Fellows, and the interested public came together at the Wissenschaftszentrum Berlin für Sozialforschung (WZB). Recent developments in artificial intelligence, such as the much-discussed capabilities and limitations of generative language models like GPT-4, have stirred controversies about how AI is and will be changing economic and societal outcomes. The design and governance of AI, the role of competition policy, and the potential need for regulatory intervention have moved into the spotlight, becoming core issues in current academic and policy discussions.

The conference panelists and participants discussed the road ahead for shaping institutions and markets to harness the benefits of AI while curbing its risks for the good of society.

In the opening session, Joanna Bryson (Hertie School of Governance), Flavio Calvino (OECD), Joshua Gans (University of Toronto), and Maximilian Kasy (Oxford University) discussed AI diffusion across firms, democratic control, regulatory trade-offs, and the economic benefits of AI. Flavio Calvino presented the OECD's current research on AI diffusion, offering insights into skill demand, patent characteristics, and the profiles of AI-adopting companies. Maximilian Kasy emphasized the need to align AI objectives — including the resolution of distributional and value conflicts — through democratic control, in order to balance the interests of those who shape AI with those affected by it. Joshua Gans opened the dialogue on AI regulation, outlining both its benefits and potential risks. He distinguished between identifiable harms, such as cheating and deep fakes, and unintended consequences, which he suggested could be uncovered through small-scale experiments. Joanna Bryson addressed the industry's capacity to adapt to regulatory interventions and provided a concluding synthesis of the arguments raised by the other panellists. This set the stage for a lively discussion, moderated by Hannes Ullrich (DIW Berlin and University of Copenhagen), on productivity gains from generative AI, security considerations, and lessons from existing digital regulations that can inform AI governance.
Watch the presentations and discussions of the panel on the shaping and governing of generative AI here.

In the afternoon session, Oren Bar-Gill (Harvard University), Joanna Bryson (Hertie School of Governance), Emilio Calvano (Luiss University), and Moritz Hardt (Max Planck Institute for Intelligent Systems) addressed the harms, challenges, and regulation of AI in consumer markets. Oren Bar-Gill introduced different categories of harm in AI-based decision-making in consumer markets, distinguishing between varying levels of market differentiation and consumer sophistication to examine algorithmic price discrimination, algorithmic targeting, and AI-induced misperception. Emilio Calvano followed with illustrative examples of how AI can capture rents through collusion or exploitation, highlighting emerging challenges such as flawed estimations of consumer preferences, a lack of transparency, and consumer manipulation. Moritz Hardt offered insights from machine learning research, discussing the performativity of predictions shaped by data steering. He argued that AI harms result from such steering, with the degree of performativity depending on the power dynamics of authority and reach. Joanna Bryson reflected on calls from EU telecommunications companies to permit large mergers and provided an overview of global market dynamics in the technology sector, supported by current data and research. The presentations were followed by a lively discussion moderated by Hans W. Friederiszick (European School of Management and Technology), in which the panelists underscored the difficulty of identifying, quantifying, and addressing harms, while also pointing to the potential benefits of AI tools for consumers.
Watch the presentations and discussions of the panel on how AI might exacerbate consumer harm in concentrated markets here.

In the closing policy roundtable, panellists Brando Benifei (European Parliament), Francesca Bria (University College London and Radiotelevisione Italiana), Alena Buyx (TU Munich), and Amba Kak (AI Now Institute and Signal Foundation), together with moderator Anna Sauerbrey (Die ZEIT), engaged in a lively debate on the state of AI policy. The speakers explained the provisions of the recently adopted AI Act, highlighted the need for publicly owned digital infrastructure, and discussed how regulatory enforcement could be strengthened.
Watch the roundtables presentations and discussion on the state of current and future policy initiatives concerning AI here.

The conference program and further information about the speakers can be found at the conference website.
 

Back