Putting Ethics First
Machine learning helps when we’re drowning in data but proactive AI ethics is needed. The MadHATTERs Editorial, 15 July 2020
MadHATTERs Weekly
Like what you read? Subscribe here.
A UNESCO-backed collaboration is turning to AI to help policymakers beat COVID-19. Launched last week, the Collective and Augmented Intelligence Against Covid-19 (CAIAC) wants to use machine learning and human expertise to make sense of the pandemic-related health, social and economic data that’s being generated globally. Expect a decision-making tool within two months that focuses on three initial areas: digital contact tracing, misinformation and the second- and third-order effects of COVID-19 on marginalised communities.
Too much data, too little time. That’s the driving force behind the CAIAC and several other initiatives trying to organise and analyse a ‘vast’ amount of raw data for use in policymaking. No surprise they’re turning to machine learning for help: speed is crucial to make sense of all this data, given the urgent need for evidence-backed policies.
But many issues abound. If the quality of the data doesn’t measure up like in the case of the UK’s coronavirus testing data, this would certainly impact decisions. And we all know how facial recognition tech is only as effective as the (biased) machine learning training behind it. The lack of ‘robust’ ethics for AI also makes things more difficult in a crisis scenario like the current pandemic. Cambridge University researchers are calling for an ‘ethics for urgency’ — a new, faster and more proactive way of doing AI ethics. For AI to be really useful in a crisis, they say, the ethics need to be already built into these systems instead of seen as an afterthought.
Putting ethics first when it comes to AI, or with anything we do in tech, is the only way we can trust the decisions that are made using them.
MadHATTERs is a weekly newsletter covering technology, personal data, and the Internet. Its perspective championing decentralised personal data is led by Dataswift with the Hub of All Things (HAT) technology. If you like what you read, subscribe to receive MadHATTERs in your inbox. Find out more about the HAT at www.hubofallthings.com