Leadership, ethics and the practical use of AI in social housing (delivered for HQN, 27th Jan 2026)

Описание: In this HQN talk, Guy Marshall explores leadership, ethics/governance, and practical AI use in social housing, with a consistent focus on one outcome: decent homes that people can afford.

He opens by mapping the “solar system” of social housing and argues that leadership is often a system property shaped heavily by UK government, regulation, vendors, and consultancies—not just housing association executives. That matters because AI will accelerate sector feedback loops (repairs → complaints → scrutiny → financial pressure), and without strong internal capability, AI direction will default to whoever is selling it.

The talk then reframes “AI governance”: most current concerns (GDPR, cybersecurity, access control, data minimisation, data quality) are foundational digital governance, not AI governance. AI-specific governance is about fairness testing, model performance, drift monitoring, accountability in procurement, meaningful human oversight, adversarial testing, workforce impact, and intersectionality-aware assessment.

Finally, Guy moves to practical delivery: adopt short learning loops (build → measure → learn), but judge success by customer outcomes, not novelty. He shares examples across the sector (analytics acceleration, Power Apps for repairs capture, accent-aware NLP for contact centres) and challenges providers to demand real evaluation metrics from vendors—especially where long-running “trials” never scale or stop.

He closes with concrete sector-wide moves: data standards (HACT's UK Housing Data Standards), shared assets, potential analytics standards, and cross-provider collaboration (including UK Hive launching Spring 2026)—because the sector’s long-term mission and recent tragedies demand better visibility into what is truly blocking performance, and disciplined use of data (and AI) to fix it.