YWA deployed Talk to the City (T3C) as one of their research platforms, introducing an innovative interface that bridged the gap between large-scale data collection and qualitative insight. The platform's key innovation lay in its interactive visualization interface, which allowed researchers to dynamically explore relationships between different data points while maintaining direct access to individual testimonies. T3C's user-centered design enabled seamless integration of 85 hour-long video interviews. This unique approach preserved the context and nuance of personal experiences while enabling pattern recognition across the broader dataset.

Technological progress is often equated with invention, yet history shows that invention alone rarely transforms society without effective diffusion. This paper examines the persistent “diffusion dilemma,” the lag between the emergence of general-purpose technologies and their broad, productive adoption. Drawing on historical examples such as the delayed uptake of tractors and industrial electrification, it argues that institutional, cultural, and infrastructural factors, rather than technical potential, govern the pace of transformation. Applying this framework to contemporary artificial intelligence, the paper contends that capability development has outpaced our mechanisms for integration and governance. It introduces the concept of wrappers, the connective layer that translates foundational models into usable, value-aligned systems, and proposes that the central challenge of innovation today lies not in creating more powerful technologies but in designing the social and organizational conditions that allow them to diffuse safely and meaningfully.

This paper examines the effects of large language models (LLMs) on UK labor market outcomes, comparing outcomes across firms and occupations from 2021 to 2025 based on their differential exposure to LLM capabilities. Highly exposed firms reduce employment, particularly in junior positions, and sharply curtail new hiring, with technical roles (software engineers, data analysts) experiencing the steepest declines. The observed effects concentrate almost entirely in high-wage segments, suggesting LLMs may compress wage inequality.

To be effective, augmented deliberation must preserve people's ability to genuinely engage with and be transformed by the deliberative process. This article proposes the Goldilocks Framework for Augmented Group Intelligence, and argues that the optimal use of AI in deliberation is about balancing between two critical elements: participants’ agency in steering and explaining deliberative outcomes, and participants’ commitment to those outcomes.

The EU's AI Act is transitioning from voluntary guidelines to enforceable standards in August 2025, with Codes of Practice serving as interim compliance measures that are technically voluntary but effectively mandatory since they grant regulatory presumption of conformity. These standardization frameworks are increasingly integrated into market mechanisms like insurance and procurement, creating a dual approach where both regulatory mandates and economic incentives drive AI risk management.

This paper argues that market governance mechanisms should be considered a key approach in the governance of artificial intelligence (AI), alongside traditional regulatory frameworks. We examine four emerging vectors of market governance, demonstrating how these mechanisms can affirm the relationship between AI risk and financial risk while addressing capital allocation inefficiencies.

This paper examines the systemic risks posed by incremental advancements in artificial intelligence, developing the concept of ‘gradual disempowerment’, in contrast to the abrupt takeover scenarios commonly discussed in AI safety.

As AI capabilities advance, we need robust methods to explore complex scenarios and their implications. This post explores when wargaming is most effective as a research tool for AI safety, and which types of problems are best suited to this methodology.