ROAR Insights Members

Insights ¦ Closing the fuel poverty gap: A plan for targeted energy support

Published by: Public First Search for original: Link Key Take Aways Persistent Crisis: Britain faces a sustained affordability crisis, with household energy bills over £700 higher than winter 2021 levels, affecting up to 6.1 million households in fuel poverty. Widening Gap: The fuel poverty gap has nearly doubled since 2020, now averaging over £400 in England, reflecting the deepening severity of the crisis. Support Expiry Risk: The current Warm Home Discount (WHD) is due to expire in April 2026, risking… Read more

Insights ¦ Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence

Published by: Stanford University and NBER Search for original: Link Key Take Aways Entry-Level Impact: The proliferation of generative AI is beginning to affect entry-level employment, particularly for workers aged 22–25 in highly AI-exposed occupations. Employment Decline: Early-career workers in the most AI-exposed roles have experienced a 13% employment decline compared to pre-adoption levels, even after accounting for firm-level shocks. Cohort Disparity: While overall employment remains strong, young workers in AI-exposed jobs have seen stagnant employment since late 2022, unlike… Read more

Insights ¦ The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity

Published by: Apple Search for original: Link Key Take Aways Current frontier Large Reasoning Models (LRMs) exhibit fundamental limitations in developing generalisable reasoning capabilities, especially beyond certain complexity thresholds. Empirical analysis reveals three distinct reasoning regimes: standard LLMs outperform LRMs at low complexity; LRMs gain advantages at medium complexity; both collapse at high complexity. Scaling propensity shows a counterintuitive pattern: reasoning effort (token usage) increases up to a threshold but then declines despite facing more complex problems. Performance collapse occurs… Read more

Insights ¦ State of AI in Business 2025

Published by: MIT NANDA Search for original: Link Key Takeaways The majority of organisations are high adopters of GenAI tools like ChatGPT and Copilot, yet only a small fraction (around 5%) are realising meaningful business impact or transformative change. The core barrier to scaling AI solutions is not model quality or regulation but the organisation’s approach — specifically, its ability to build systems that learn, adapt, and retain context over time. Big firms lead in the number of AI pilots… Read more

Insights ¦ Evolving our funding model

Published by: Financial Ombudsman Service Search for original: Link Key Take Aways The Financial Ombudsman Service is evolving its funding model to better reflect case workload, aligning with principles of fairness, simplicity, and sustainability. The consultation proposes further differentiation of case fees based on case stage and outcome, moving beyond the existing flat fee structure introduced with professional representatives. Differentiation by case stage aims to incentivise early resolution, with lower fees at initial stages and higher fees for cases requiring… Read more

Insights ¦ Call for Input Modernising the Redress System

Published by: Financial Conduct Authority Search for original: Link Key Take Aways The current redress framework effectively handles individual complaints but struggles with mass redress events involving large volumes of complaints and systemic harm. Increasing complaint volumes, particularly from professional representatives, can overwhelm firms and the Financial Ombudsman, leading to delays, increased costs, and potential disorderly firm failures. Greater collaboration between the FCA and Financial Ombudsman is planned to improve early identification and management of issues with wider implications (WWI),… Read more