Anjali Mazumder, artificial intelligence, justice and human rights lead at the Alan Turing Institute, works across three programmes; the AI programme which focuses on safe and ethical AI, the defence and security programme and the public policy programme.
She explained that one of her areas of focus is the way in which data science and artificial intelligence can be used to combat modern slavery and other exploitive crimes. She said: “Generally we are looking at modelling pathways of exploitation and use these to determine where we could intervene along the pathway to either prevent or protect [people] from harms.”
To give background to the problem, Mazumder stated that there are estimated to be 40 million individuals in slavery at present. In addition, there is the Sustainable Development Goal 8.7 which calls for the eradication of forced labour, modern slavery and human trafficking by 2025. She pointed out that if this goal is to be met, 9,000 people a day will have to be brought out of slavery, which may be overly optimistic.
In terms of the data on the estimated number of people held in slavery, Mazumder said that the number came from various sources and there is acceptance of the fact that not all victims are being counted because they use ‘multiple disparate sources of data with large gaps.”
Labour exploitation at sea as well as domestic servitude, as depicted in the film ‘I am slave’, keep people well-hidden and much harder to count. “There are lots of people that we may not be reaching or understanding. So we are looking at other ways to capture this data and identify victims as well as perpetrators, whether they be recruiters, users or abusers,” she said.
She spoke of Slavery from Space project from The Rights Lab of the University of Nottingham as an example of an innovative way to find hidden victims of slavery using satellite images. Mazumder said: “The idea is that identifying brick kilns in India in rural areas is a potential indicator of a risk of bonded labour.”
Satellite imagery data is one source but there is also mobile data and transactional data, and when those sources are integrated some individuals may be missed while others may be double counted.
Another challenge is the changing nature of modern slavery and human trafficking and so those fighting it with data need to keep pace with the changing patterns of exploitation.
Going forward, Mazumder said she will be looking at ways to develop more tailored interventions to suit survivors. “One of the things I think is quite crucial in the space is taking a multi-sector and multi-disciplinary approach,” she said.
She illustrated that she and her peers at the Alan Turing Institute have initiated this by leading a proposal of an AI global working group on modern slavery and human trafficking. The first meeting has already taken place in February 2019 and comprises organisations including the Alan Turing Institute, United Nations University Centre for Policy Research, Tech Against Trafficking, University of Nottingham Rights Lab, Arizona State University Global Security Initiative.
Anjali Mazumder was speaking at Big Data, AI and the future of crime and justice hosted by the University of East London in association with the British Society of Criminology, Crime and Justice Statistics Network.