.png?width=200&height=40&name=Risk%20Insights%20-%20Logo%20-%20(RGB%2c%20Horizontal%2c%20Small).png){kind=small}
TL;DR • "Basic" machine learning amplifies familiar risks (e.g. security, privacy). • But even these...
TL;DR • This resource is from ICA (Insurance Council of Australia) and CSIRO (Australia's national s...
TL;DR • Some decisions need precision (e.g., safety, fairness). • Others just need "good enough". • ...
TL;DR • Passive trust often creeps in when human oversight becomes routine and passive. • Regular au...
TL;DR • Demographic survey data can help inform public education and intervention, but isn’t reliabl...
TL;DR • This resource is from IAIS (the International Association of Insurance Supervisors). • The f...
TL;DR • Good design makes it easier to read and understand dashboards and visuals. • Keep visuals cl...
TL;DR • More data means more potential patterns. • Correlation can reflect causation, but isn’t proo...
TL;DR • This resource is from NIST (the U.S. National Institute of Standards and Technology). • A gl...
TL;DR • Don’t default to a “standard” or “comprehensive” package. • Start with a simple top-down foc...
TL;DR • Acceptable Use Policies are typically too long and complex for most people to actually read....
TL;DR • Some countries let insurers use sex for pricing/underwriting (as an exemption to discriminat...
TL;DR • A detour from our usual focus. • AI-writing detectors can wrongly label genuine work as AI-g...
TL;DR • Unlike other discrimination categories, Age has/is a single attribute. • It's often visible ...
TL;DR • Disability has 9 key attributes. • Some are visible in data, but most are inferred or hidden...
.jpg)
TL;DR • Sex has eight key attributes, some direct in data, but mostly inferred. • Many attributes ca...
TL;DR • Race has four/five key attributes. • Some racial attributes are easier to spot and manage th...
TL;DR • Protected attributes hide in unstructured data, proxies, and third-party sources. • Knowing ...
TL;DR • A management system standard: covers governance processes, but doesn't ensure integrity or f...
TL;DR • Narrowly applied AI definitions miss the point. • Robodebt showed us that a "simple" system ...