Modern Slavery and AI: An Introduction to where those two Worlds can Meet

Regardless of your domain or interests, you have probably heard about two terms that have recently been into the spotlight, “artificial intelligence”, and more recently, “modern slavery”. How can they be linked?

While it would be very interesting to assess how artificial intelligence (AI) can contribute to the spread of modern slavery, this article is looking at how AI is helping fight against modern slavery.

There are many applications of AI for Good, but where does the fight against modern slavery fit in the 2030 Agenda for Sustainable Development?

It is under the United Nations’ SDG 8, target 8.7: calling upon states to “Take immediate and effective measures to eradicate forced labour, end modern slavery and human trafficking and secure the prohibition and elimination of the worst forms of child labour, including recruitment and use of child soldiers, and by 2025 end child labour in all its forms”

With less than ten years to go, to completely eradicate this crime in all its forms, we need to understand how does slavery look today, how can stakeholders accelerate their work to eradicate it, and how AI can be of assistance?

This article presents an introduction to the problem of modern slavery and shows the key role of different stakeholders in the fight against it. For each of them, the article would present some of the most promising AI applications that can assist them in this journey.

If you have not yet heard the term ‘modern slavery’, you have probably heard of ‘human trafficking’. While widely more recognized, the term ‘human trafficking’ is considered a subset of the modern slavery phenomenon, arguably to suffice to capture its complexity.

Resolving the issue of modern slavery requires agreeing upon a definition. As this is not the case yet, it is important to refer back to the available legal frameworks and definitions to determine what modern slavery is. For example, when the new Modern Slavery Act (MSA) Australia was constructed, the definition of modern slavery was built informed by the existing legal frameworks. The Australian Modern Slavery Act 2018 is the first national legislation in the world to define modern slavery.

The term ‘modern slavery’ is used to describe situations where coercion, threats, or deception are used to exploit victims and undermine or deprive them of their freedom. Modern slavery is referred to as an umbrella term that encompasses “any situation of exploitation that a person cannot refuse or leave because of threats, violence, coercion, deception, and abuse of power. “ Those instances can be:

5 Common forms of modern slavery in our world today.  Picture source: End Slavery Now

Note: This article uses both ‘modern slavery’ and ‘human trafficking’ to refer to this crime.

While measuring the hidden crime of modern slavery is not an easy task, Walk Free Foundation and the International Labour Organisation(ILO) have been publishing estimates of slavery in the Global Slavery Index.  The most recent edition shows that there are approximately 40.3 million enslaved in the world, out of which, one in four of the victims is a child. What is more, women and girls are disproportionately affected. The results also show how modern slavery is not an ‘over there problem’, but it happens in every country of the world.

Modern slavery continues to proliferate due to a lack of market pressures and regulations. Modern slavery is the failure of the market with very few regulations in this regard. This crime is the third-largest in the world, after drug and counterfeit goods smuggling, generating an estimated $150 billion of illegal profit. It is now cheaper than ever to trade in human beings.

“People do not enslave people to be mean to them; they do it to make a profit.” – Kevin Bales

Modern slavery is the exploitation of vulnerabilities. It thrives on the exploitation of the most in the need of protection putting children, migrants and refugees, at the top of the most likely to become victims.

Slavery flourishes where there is a weak rule of law, poverty, discrimination against marginalized groups, instability brought on by war, conflict or natural disasters, etc. In the face of the population boom and the historic flux of people, vulnerable people who are looking for stability and safety, in the absence of alternatives, fall prey to the false offers of perpetrators.

When this epidemic meets a global pandemic, the need to discuss and address this issue could not be more urgent. COVID-19 has accelerated the vulnerability not only of those already exploited and survivors, but it has been expanding those vulnerabilities to those who were just meeting the threshold of safety and security. Furthermore, the existing efforts to fight modern slavery are constantly hampered.

In the last few months, investigations show how the pandemic has dramatically increased the risk of exploitation. With the global lockdown and job losses, migrant workers were left without security and safety, unprotected not only for the disease but also for the exploitation of modern slavery perpetrators.

What is more, the reduced consumer demands, obstructed supply chains, and logistics, and factories’ shutdowns, increased the challenges for those who were already working at the bridge of exploitation, adding them to the list of possible victims.

Yet, this increase in chances of exploitations has been seen also in the type of modern slavery more commonly found in the developed world, such as online sexual exploitation of women and children. With the shift towards more online activity, and without an infrastructure ready to accommodate for those influxes, vulnerable women and children are facing a higher chance of online recruitment and exploitation. Moreover, being in lockdown reduces the possibility of reporting, as the victims are finding fewer chances to report their abusers.

“Tragically, we are finding many traffickers/pimps are taking advantage of individual’s vulnerabilities during these tragic times. Myself and other leaders in the anti-human trafficking space, including the FBI, are seeing that right now is a time of increased recruiting and grooming of victims via online connections.” – Tomas J. LaresPresident/CVO United Abolitionists

Fighting modern slavery is our collective responsibility.  No stakeholder or individual alone can solve this issue. Collaboration and coordination among stakeholders and initiatives are key to eradicate modern slavery and technology can assist this journey.

Before analyzing how technology can help in tackling slavery, it is important to acknowledge how the business of modern slavery is facilitated by the unregulated technological and online environment.

Criminals have been enabled to grow their business and increase abuse by misusing the technology. Under the recent study “Leveraging innovation to fight trafficking in human beings: A comprehensive analysis of technology tools” conducted by the Organization for Security and Co-operation in Europe (OSCE) and Tech Against Trafficking, lists the ways in which the criminals are taking full advantages of the benefits brought by technology:

For all these examples presented, there are opportunities to utilise technological solutions designed to identify, track and ultimately eliminate the abuse. However, the same study also shows that unfortunately, as of December 2019 the number and the scale of technology initiatives and tools designed to fight against modern slavery do not match the size of the problem.

Therefore, it is imperative to accelerate and scale the existing work against modern slavery and leverage technology, especially AI to do so.

“Imagine the techniques that Google and Facebook are using to make tons of money—understanding people the way they connect, what their interests are, what they might buy or the activities they engage in,”. “We can apply those same techniques—data mining, text mining, what’s called graph mining—AI that’s being used for legitimate and really profitable purposes, to track these illicit behaviors.” – Professor Dan Lopresti

Picture source: https://www.telegraph.co.uk/

It is time to turn the technology used to exploit people towards saving them.

Note: As with the term ‘modern slavery’, there is no communly agreed definition for ‘artificial intelligence’ (AI) either. For the purpose of this paper, the definition used will be the one shaped by The Future Society:  AI = Big data driven, Machine learning algorithm-centric, Socio-technical systems Powered by supercomputing.

During the last few years, there has been huge excitement about the potential use of AI for social good. While  AI presents huge opportunities to accelerate our work towards meeting Agenda 2030, AI is not a silver bullet. The deployment of AI is complex and strongly dependent on data.

Yet encouraging developments which leverage technological innovation, allows for the footprints left by the perpetrators, especially in the online sphere, to be traced and transformed in valid data to train AI algorithms to fight against modern slavery,

“If you are going to traffick humans, there needs to be a retail component of what you are doing. There needs to be a mechanism for traffickers to put up ads and make known that they have humans for sale. This is an opportunity where you have public facing data in web pages, on social media, in ads, in chats, etc. This has to be publicly available to facilitate purchase.” – Dr. Gary M. Shiffman

Recently very promising initiatives have been put forward illustrating the potential of AI against modern slavery. Conferences such as Code 8.7, in February 2019, which brought together more than 120 people from the computational research and artificial intelligence communities with those working to achieve Target 8.7 as well as the recent mapping exercise done by OSCE and Tech Against Trafficking, help us identify and present the most promising AI applications against modern slavery.

Each stakeholder has a key role to play in eradicating modern slavery, and AI is already here to assist them.

Governments have the central role to lead the national and international response to modern slavery. This crime is fully benefiting from the governance gaps that allow the criminals to perpetuate it with impunity. The governments should set standards through legislation and support businesses and civil society to act against modern slavery. Once passed, those legislations need to be closely monitored and evaluated under strict deadlines. This process can be efficiently facilitated with the assistance of AI.

For example, since 2010, with the pass of California’s Transparency in Supply Chains Act, transparency in supply chains legislation started to appear, asking businesses to analyze their supply chains and report in annual statements how they are addressing modern slavery. Since then, governments started to pass similar legislation at a national level, the first being the UK, with the Modern Slavery Act(MSA) of 2015, and more recently the Australian Modern Slavery Act of 2018.

With thousands of companies publishing reports each year, governments and researchers lack the capabilities to manage and analyse a large number of reports. What is more, other governments are working towards passing similar legislation, and thus, a more sustainable and scalable solution is required. AI can be part of the answer.

Law enforcement agencies are also facing an uphill battle to conduct complex investigations and respond, prevent, protect, and prosecute, in time to piece together data from a multitude of sources. The complexity of the criminal organizations is also expanding, and thus, AI has a huge potential to assist with the parts of the analysis and investigation that can be automatized.

For instance, the use of natural language techniques, text mining, and data mining techniques will allow law enforcement to identify patterns of behaviour that would reflect illegal activities related to modern slavery. AI can assist the law enforcement agencies to put together the pieces to uncover the big picture of a criminal organization’s activities as well as to pinpoint leads that will enable them to follow individual leads. Also, the transactions associated with the crime of modern slavery can be traced using AI, looking across sparse data streams to create the evidence needed for investigation and persecution.

Examples of AI driven solutions

Project AIMS (Artificial Intelligence against Modern Slavery): develops an AI solution to read and benchmark the statements produced by companies under the UK Modern Slavery Act. It uses Natural language processing(NLP) and computational linguistics to check the quality of companies’ reporting. This tool will be open sourced and could be further developed to fit the needs of the Australian MSA reports and any other non-financial reporting requirements.

Magnet.AI, is helping to identify and apprehend child predators. In particular, Magnet.AI uses machine learning to narrow massive amounts of content. This improves the triage process by helping forensic examiners to focus their efforts more accurately, saving time and reducing the risk of human error in missing messages.

Traffic Jam uses AI to help law enforcement find victims and enable them to take down organized criminal networks. In 2019, Traffic Jam was used to identify an estimated 3,800 victims of sex trafficking. Traffic Jam is used by law enforcement agencies in the United States, Canada, and the United Kingdom.

Childsafe.ai is an artificial intelligence platform for monitoring, graphing and modelling child exploitation risk on the Web. In use by law enforcement agencies around the United States, childsafe.ai actively collects signal of exploitation threats from online ecosystems where they are known to occur, modelling that signal into probable risk.

On their side, businesses have a key role in eradicating modern slavery, being uniquely placed to fight this abuse nationally and internationally through their complex supply chains.  Many business leaders agree that their organizations should and do respect human rights, implement due diligence, and create change across borders through their transparency and action on their supply chains.

However, those extremely complex and opaque supply chain structures provide challenges associated with assessing, identifying, and managing the links of the chain at risk of being plagues with modern slavery. Yet, they agree that their responsibility and interest is to eliminate modern slavery from their supply chains, not only to meet the requirements of the legislation, where they exist but also to ensure the long-term sustainability and survival of their business.

The pressure is exercised on the companies from better-informed consumers, legislations as well as investors who are starting to include the modern slavery risks in their processes and conditionalities.

AI can be utilised to assist companies to identify and analyse their risks in supply chains and enhance due diligence. Connecting ambiguous and sparse datasets, AI can also help companies to spot deception and exploitation in acquiring legitimate assets or placing victims within legitimate employment. AI can scale and progress the supply chain data analysis, connecting dots with existing data (on areas, products, organisations, individuals, accounts, tax numbers, etc) that can flag risk for the company. The evidence and literature in this domain have been flourishing for the last few years, and AI can help companies integrate this knowledge into their assessment of their supply chains.

Examples of AI driven solutions:

Bext360 utilizes machine vision and artificial intelligence (AI) to source high-quality agricultural products directly from the source. Bext360 makes digital/mobile payments to the farmers, communities, banks, and other stakeholders utilizing blockchain payments directly to the stakeholders. The bext360 system then tracks the goods from the source to the end consumer, allowing the end consumer to interact directly with the communities that provide the goods.

DDIQ developed by Exiger, is an AI-based automated due diligence solution that accelerates and enhances risk assessment related to clients, investments, transactions, third-parties and counterparties.

Phylagen combines the most advanced DNA sequencing technology with artificial intelligence to compare these fingerprints and verify the origin of products across global supply chains.

“Human trafficking is a very hard crime to prove because you have to pull together a lot of pieces of information,”. – Professor Dan Lopresti

Building this bigger picture on modern slavery would be impossible without the key input from the civil society, academia and individuals.

Civil society and consumers can promote best practices, raise awareness, and influence businesses and government behaviours. NGOs can engage with other stakeholders to share their expertise, data, and enhance accountability.

What is more, academia has a crucial role in advancing research and building evidence that can be used to inform policy, legislation, and action.

Looking at the future, that solutions for modern slavery should be sustainable, and those saved from slavery to be protected against further exploitation. To generate a comprehensive strategy against modern slavery, the voice of the victims should be integrated through its design and execution, aiming at preventing, identifying and protecting. All those stakeholders play a key role in either designing, developing and deploying AI solutions to create global impact against modern slavery.

Examples of AI driven solutions:

Traffik Analysis Hub allows analysts to draw from a rich pool of data and more easily identify trafficking patterns, networks and hotspots. Using IBM Watson – AI, machine learning, and natural language recognition – an intelligent “golden tagging” schema is applied to the data. In addition, the TA Hub solution is trained to recognize terms and incidents related to human trafficking in the unstructured content and structure it along with the golden tags schema.

The University of Nottingham, UK has applied artificial intelligence to satellite imaging to map forced labour patterns and predictions. An example is mapping brick kilns in South Asia, which have proven in the past to be hotspots for forced labour and debt bondage.

Trace-the-Face program uses photos to match up missing family members who have been separated due to migration and conflict. Trace-the-Face has been updated and extended by Microsoft to perform automatic face detection and matching using machine learning, via Microsoft Cognitive Services Face API.

Crisp Crisis Monitoring: Crisp was established to protect children and teenagers using online games and social networks from abuse, sexual exploitation, cyberbullying and other online threats. Crisp’s extended intelligence crawls thousands of public and dark web data sources to deliver Social Media Safety for organisations and users.

Modern slavery is a result of the exploitation of those most vulnerable and of our collective failure to protect them. Modern slavery touches all of us across the globe, through our food, clothes, electronics, etc. It is our responsibility to eradicate it once and for all, not just in law but also in practice.

Modern slavery is a global crime of proportion. While technology is facilitating this crime, it is also part of the solution. We need to deploy solutions at scale, to catch up and win this fight. To do so, AI is best placed to assist, but only when designed ethically and with the wellbeing of the victims in mind.

Those solutions shall be designed with a multistakeholder and multidisciplinary approach. What is more, it is time to have a look at the existing technologies and increase awareness of what is available to avoid duplication and fragmented and disjointed development. With less than ten years to achieve Target 8.7, coordination and innovation are key.

For more examples of AI and other technology solutions against modern slavery, check out Code 8.7 and Tech Against Trafficking’s Interactive Map.