Regulating for algorithm accountability: global trajectories, proposals and risks

The extensive use of algorithmic decision-making in all domains of social life requires specific accountability mechanisms and regulations that ensure meaningful redress. This is an especially hard task when little information on the algorithms in use is available in the public domain and their implementation rational, and the organisations ultimately responsible for their functioning, remain opaque.

While there is little consensus over what approach to take, countries across the world have started designing and applying different mechanisms to boost algorithm accountability.

In 2018, New York City launched a Task Force to make recommendations on how the city should manage automated decision-making systems. Earlier this year, New Zealand issued an Algorithmic Charter to be deployed in case of high-risk applications. Canada has developed a model of Algorithmic Impact Assessment, a scorecard that helps identify the level of risk of an algorithm and mitigation factors. More recently, the UK Institute for the Future of Work’s Equality Task Force has released a report highlighting gaps in legal protection and mechanisms for accountability, and calling for new legislation: an Accountability for Algorithms Act.

The Ada Lovelace Institute, the Institute for the Future of Work and international experts in algorithm accountability surface key concerns and relate them to the governance and regulatory landscapes of different national contexts. In this event we ask:

• How do we ensure that algorithmic systems and the agencies and organisations deploying them are truly accountable? Is new regulation necessary?
• What can we learn from the different approaches in New Zealand, Canada and New York City?
• How do they relate to their respective regulatory and administrative contexts?

Image credit: your