Bridging the Gaps: Human Trafficking Data in the Global South
Note: The following is a distilled, blog-version of the presentation I recently gave at the inaugural Human Trafficking Data Conference at Southern Methodist University, in Dallas. You can also view the full presentation slides below.
Reflecting on my decade-long journey in the anti-trafficking field, I've encountered countless stories of resilience and tragedy that highlight the complexities of human trafficking in different cultural contexts. Working in countries like Cambodia, Thailand, and the Philippines has exposed me to the nuanced challenges faced by survivors, service providers, and grassroots organizations.
The Data Crisis: Unraveling the Systemic Barriers
The data crisis in the Global South is a stark reality that hinders progress in the fight against human trafficking. This crisis is deeply rooted in the region's colonial past and the persistent power imbalances that shape its present. As Western nations and international organizations continue to exert influence over the Global South through development aid, economic policies, and governance structures, the legacies of colonialism and the neocolonial system remain entrenched.
These historical and systemic factors have significant implications for how we understand and address trafficking in the region. Traditional data collection methods, such as surveys conducted by the International Labour Organization (ILO), the International Organization for Migration (IOM), and UNICEF, often fail to capture the full scope and complexity of the problem.
ILO surveys, for example, primarily focus on trafficking in formal labor markets, potentially overlooking the vast informal sectors where exploitation is rampant. IOM surveys, on the other hand, adopt a migrant-centric approach, which may not fully capture the experiences of non-migrant populations, including those affected by domestic trafficking. UNICEF surveys, while crucial for understanding child trafficking, may have limited scope and face accessibility challenges in remote or conflict-affected areas.
Moreover, these surveys often rely on formal reporting mechanisms and official records, leading to significant underreporting due to fear, stigma, and lack of trust in authorities. As a result, the true scale and nature of trafficking in the Global South remain obscured, with estimates suggesting that only a fraction of cases are ever officially documented.
The consequences of these data gaps are severe, hindering policymaking, resource allocation, progress tracking, and accountability. Without accurate, comprehensive data, policymakers and practitioners are left to make decisions based on incomplete or distorted information, leading to a mismatch between the needs of survivors and the resources allocated to support them.
As someone who has witnessed firsthand the challenges posed by the neocolonial system in my work in the Philippines, I know that confronting the data crisis requires us to grapple with the complex web of factors that perpetuate it, including the historical legacies of colonialism, the persistent inequalities in global power structures, and the limitations of traditional research methods.
But I also know that change is possible – and necessary. By embracing participatory data practices, harnessing the power of ethical AI, and working towards new governance models that center the voices and experiences of survivors and communities, we can begin to bridge the data gaps and build a more effective, equitable, and sustainable response to trafficking in the Global South. The path ahead may be challenging, but the stakes are too high for us to shy away from this critical work.
The Power and Potential of Ethical AI
Alongside participatory data practices, the rapid advancements in artificial intelligence (AI) and machine learning are opening up new frontiers in the fight against trafficking. In particular, the development of large language models (LLMs) – AI systems trained on vast amounts of text data – has the potential to revolutionize how we understand and respond to this complex issue.
LLMs like those developed by Anthropic (Claude) and Inflection (Pi) have the capability to process and analyze massive amounts of unstructured data, from survivor testimonies and case reports to academic literature and social media posts. By identifying patterns, connections, and insights that may be difficult for human analysts to discern, these models can help us to build a more comprehensive, nuanced understanding of trafficking. For example:
LLMs could be used to analyze survivor narratives at scale, identifying common risk factors, recruitment tactics, and methods of control used by traffickers. This information could then be used to inform more targeted, effective prevention and intervention strategies.
LLMs could be trained to identify patterns of trafficking activity across different data sources, helping to map hotspots and track the movement of victims and perpetrators.
However, as with any powerful technology, the use of AI in anti-trafficking efforts raises important ethical considerations:
Bias: Ensuring that AI systems are not perpetuating or amplifying existing biases in the data or society.
Fairness: Guaranteeing that the benefits and risks of AI are distributed equitably across different populations and stakeholders.
Transparency: Ensuring that the development and deployment of AI systems are transparent and understandable to those affected by their use.
Accountability: Establishing clear mechanisms for holding AI developers and users accountable for the impacts of their systems.
The development and deployment of AI systems in the fight against trafficking must be guided by robust ethical frameworks and principles. Models like Claude, which employs constitutional AI techniques to align its outputs with predefined values and constraints, offer a promising approach to embedding ethics into the very architecture of these systems.
Ethical AI also demands ongoing monitoring, evaluation, and adjustment to ensure that these systems are not perpetuating or amplifying existing biases and power imbalances. This requires active collaboration between AI developers, anti-trafficking experts, and affected communities to ensure that these tools are being developed and used in ways that align with the needs, values, and priorities of those they are intended to serve.
Ultimately, the responsible and ethical use of AI in anti-trafficking efforts has the potential to significantly advance our understanding of this urgent issue and drive more effective, targeted interventions. By leveraging the power of these technologies to identify patterns and generate insights at a scale and speed that would be impossible through manual analysis alone, we can accelerate progress towards a world free from trafficking and exploitation.
However, realizing this potential will require more than just technical innovation – it will require a deep commitment to ethical principles, participatory approaches, and equitable partnerships. It will require us to grapple with the complex power dynamics and structural inequities that shape the landscape of trafficking and anti-trafficking efforts, and to work towards more just, inclusive models of data governance and technology development.
Real-World Applications and Challenges
As we explore the potential of participatory data practices and ethical AI in the fight against trafficking, it's important to ground our discussion in real-world applications and challenges. While these approaches offer immense promise, their implementation requires careful consideration of the complex realities and constraints faced by anti-trafficking actors on the ground.
One area where participatory data practices are already showing promise is in the realm of qualitative research. By involving survivors and community members in the design, collection, and analysis of qualitative data – such as interviews, focus groups, and storytelling initiatives – researchers can gain a more nuanced, contextually grounded understanding of trafficking experiences and risk factors.
Participatory approaches can also help to address some of the power imbalances and ethical concerns that often arise in traditional research settings. For example, by training survivors as peer researchers or using collaborative data analysis techniques, participatory methods can help to ensure that the insights generated are not only rigorous but also reflect the priorities and perspectives of those most affected by trafficking.
However, implementing participatory research at scale can be challenging, particularly in resource-constrained settings. Ensuring the safety, privacy, and well-being of participants requires robust ethical protocols and ongoing support, which can be difficult to sustain without adequate funding and institutional buy-in. The main challenges include:
Ethical protocols
Language barriers
Cultural differences
Mistrust of outside researchers
Ethical AI, too, holds immense potential for advancing anti-trafficking efforts – but not without its own set of challenges and considerations. As mentioned earlier, AI systems like large language models can help us to analyze vast amounts of unstructured data, identifying patterns and generating insights that could inform more targeted, effective interventions.
For example, AI could be used to analyze case management data across different anti-trafficking organizations, identifying common challenges, promising practices, and opportunities for collaboration and learning. Predictive modeling techniques could also be used to identify individuals or communities at high risk of trafficking, enabling more proactive, preventative approaches.
However, the use of AI in the anti-trafficking space also raises important questions about data privacy, security, and ownership. Given the sensitive nature of trafficking data and the vulnerability of the populations involved, it is essential that any AI systems used in this context are designed with robust safeguards and protocols for protecting individual privacy and preventing unauthorized access or misuse of data.
Here's a specific example of how data privacy could be protected in an anti-trafficking AI system:
Implementing secure, encrypted data storage and transmission protocols to prevent unauthorized access to sensitive information
Using anonymization techniques to remove personally identifiable information from datasets before they are used to train AI models
Establishing clear data governance policies and processes that specify who has access to what data, for what purposes, and under what conditions
Regularly auditing and testing AI systems to identify and mitigate any potential privacy risks or vulnerabilities
There are also challenges around data quality, bias, and representativeness that must be addressed if AI is to be used effectively and ethically in anti-trafficking efforts. If the data used to train AI systems is incomplete, biased, or not representative of the full spectrum of trafficking experiences, the insights generated may be skewed or even harmful.
Overcoming these challenges will require not only technical solutions but also the development of new partnerships, governance models, and capacity-building initiatives to ensure that the benefits of AI are shared equitably and that its risks are mitigated effectively. It will require ongoing dialogue and collaboration between anti-trafficking stakeholders, AI experts, and affected communities to ensure that these tools are developed and deployed in ways that align with the values and priorities of those they are intended to serve.
Despite these challenges, the potential of participatory data practices and ethical AI to transform anti-trafficking efforts is too significant to ignore. By grounding these approaches in the lived realities of survivors and communities, and by working to address the power imbalances and inequities that shape the trafficking landscape, we can harness the power of data and technology to drive more effective, sustainable change.
New Governance Models
As we work to realize the potential of participatory data practices and ethical AI in the fight against trafficking, it is clear that new governance models will be needed to ensure the responsible, equitable use of these powerful tools.
Traditional, top-down approaches to data governance – in which data is controlled by a few powerful actors and used primarily for surveillance, profit, or PR – are ill-suited to the complex, high-stakes realities of the anti-trafficking space. Instead, we need governance models that prioritize transparency, accountability, and the active participation of affected communities in decision-making processes.
One promising approach is the development of data trusts or data cooperatives – legal structures that allow data to be pooled and managed by an independent third party on behalf of a community or group of stakeholders.
Key features and benefits of data trusts:
Enable secure, ethical data sharing among multiple parties
Ensure that data is used in ways that align with the interests and values of the communities it represents
Provide a mechanism for collective bargaining and benefit-sharing around data use
In the context of anti-trafficking efforts, a data trust could enable survivors, service providers, and researchers to share data securely and ethically, while maintaining control over how that data is used and ensuring that the benefits are distributed equitably.
For example, a data trust could be established to support collaborative research on trafficking risk factors and vulnerabilities, with clear protocols for data access, use, and sharing that prioritize the privacy and well-being of survivors. The trust could be governed by a board of stakeholders, including survivors, community leaders, and anti-trafficking experts, who would be responsible for setting strategic priorities, approving research proposals, and ensuring that the insights generated are used to inform more effective, survivor-centered interventions.
Another potential model is the use of smart contracts and blockchain technologies to enable more secure, transparent data sharing and collaboration.
Key features and benefits of smart contracts:
Self-executing digital contracts that automatically enforce the terms of an agreement
Provide a secure, tamper-proof record of data sharing agreements
Enable automated enforcement of data use policies and benefit-sharing arrangements
In the anti-trafficking context, smart contracts could be used to facilitate data sharing between organizations while ensuring that the data is used only for approved purposes and that the benefits are shared fairly.
For example, a smart contract could be set up between an anti-trafficking NGO and a research institution, specifying the terms of data access and use, the responsibilities of each party, and the mechanisms for distributing any benefits or insights generated from the research. The contract could be stored on a blockchain, providing a secure, tamper-proof record of the agreement and enabling automatic enforcement of its terms.
These are just a few examples of the new governance models that could be developed to support the responsible use of participatory data practices and ethical AI in anti-trafficking efforts. Other potential approaches include citizen juries, participatory budgeting, and community-led data stewardship initiatives.
The Bigger Picture: Recommendations for a Holistic Approach
Central to this approach is the need to examine and adapt the standard tools and methods used in anti-trafficking research and data collection. While surveys, interviews, and other traditional research methods can provide valuable insights, they often fail to capture the full complexity of trafficking experiences or to meaningfully engage affected communities in the research process.
Participatory methods such as action research, community mapping, and digital storytelling offer promising alternatives, but they require significant investment in capacity building, relationship building, and institutional change to be implemented effectively and ethically. Anti-trafficking organizations and funders must be willing to prioritize and resource these approaches, recognizing that they may require longer timeframes, more flexible funding models, and a greater tolerance for uncertainty and iteration than traditional research methods.
Key recommendations for a holistic approach:
Foster meaningful community engagement and trust building throughout the research and data collection process
Invest in capacity building and training for anti-trafficking organizations, researchers, and community members
Pilot and scale new approaches to data and technology in a responsible, inclusive way
Develop intentional feedback loops and spaces for collective learning and adaptation
Equally important is the need to foster meaningful community engagement and trust building throughout the research and data collection process. This requires moving beyond tokenistic or extractive forms of participation towards more equitable, reciprocal partnerships that center the agency and expertise of affected communities.
Practical strategies for building trust and engagement include:
Involving community members as co-researchers and decision-makers
Providing accessible and culturally relevant information about the research process and its potential benefits and risks
Ensuring that the insights generated are shared back with communities in ways that are useful and actionable for them
Capacity building and training for anti-trafficking organizations, researchers, and community members are also critical for advancing participatory data practices and ethical AI. This includes not only technical skills related to data collection, analysis, and management but also soft skills such as facilitation, active listening, and collaborative problem-solving.
Investments in training and capacity building should prioritize the needs and priorities of local organizations and communities, rather than imposing top-down, one-size-fits-all approaches. This may require adapting existing training materials and methodologies to local contexts and languages, as well as creating new resources and support systems that are tailored to the unique challenges and opportunities faced by anti-trafficking actors in the Global South.
Another key recommendation is the need to pilot and scale new approaches to data and technology in a responsible, inclusive way. This means starting small, testing and refining new tools and methodologies in collaboration with affected communities, and building in mechanisms for ongoing learning and adaptation.
It also means being intentional about the potential risks and unintended consequences of new technologies and putting in place robust safeguards and accountability measures to mitigate these risks. This may require developing new ethical frameworks and guidelines for the use of AI and other emerging technologies in the anti-trafficking space, as well as creating mechanisms for independent auditing, monitoring, and redress.
Finally, a holistic approach to participatory data practices and ethical AI must include the development of intentional feedback loops and spaces for collective learning and adaptation. This means creating opportunities for anti-trafficking stakeholders to come together to share experiences, challenges, and promising practices, and to engage in joint problem-solving and strategy development.
It also means investing in the documentation and dissemination of learning, so that promising approaches can be scaled and adapted to new contexts, and so that the field as a whole can continue to evolve and improve over time. This may require new forms of collaboration and knowledge sharing across sectors, disciplines, and geographies, as well as new incentives and reward structures that prioritize learning and adaptation over short-term outputs or outcomes.
Ultimately, the success of participatory data practices and ethical AI in the fight against trafficking will depend not only on the creativity and commitment of individual actors but also on the collective will and capacity of the anti-trafficking field to embrace new ways of working and thinking. It will require a fundamental shift in the power dynamics and incentive structures that shape the field, towards more equitable, inclusive, and accountable approaches that center the voices and priorities of those most affected by trafficking.
While the challenges are significant, the potential benefits are too great to ignore. By embracing a holistic, systems-level approach to participatory data practices and ethical AI, we can unlock new insights, innovations, and solutions that have the power to transform the fight against trafficking and advance the rights and freedoms of all people, everywhere.
