The COVID-19 crisis presents an unprecedented opportunity to leverage AI for global benefit. However, rapidly scaling up the use of AI and data to address the current crisis carries its own risks. The urgency with which new technologies must be deployed does not justify neglecting ethical oversight and risk assessment, but rather means it is crucial to find ways to do ethics with equal urgency.
In an article just released in Naure Machine Intellignece, Leverhulme CFI and CSER researchers, Asaf Tzachor, Jess Whittlestone, Lalitha Sundaram, and Seán Ó hÉigeartaigh suggest that ethics with urgency should at minimum include rapid foresight processes to anticipate potential risks, processes to ensure AI systems behave as intended, and ongoing, independent, ethical oversight. Far from slowing down deployment, we suggest that these processes could enable crucial new technologies to reap benefits more quickly, by strengthening public trust in their use.
In a follow-up interview, co-author Jess Whittlestone speaks with Douglas Heaven at the MIT Technology Review about a new ethics for urgency.