Electronic Monitoring Policy at Sea Beyond “Human Bycatch”

PROBLEM

Concern for the human rights implications of ocean management is on the rise. ¹ Notably, target 14.4 of the United Nations Sustainable Development Goals (SDGs) is to “end” illegal, unreported, and unregulated (IUU) ² fishing, which is broadly defined as fishing activities that disregard national or international laws and regulations and is linked to human rights violations. ³ Target 14.4 of the SDGs has relevance to the human rights dimensions of ocean management as its aims are to uphold the sustainable management of marine resources, which IUU threatens when vessels contribute to overfishing and the depletion of fish stocks, but also when fisheries laborers are exploited. ⁴ At the same time, fisheries production largely remains an “obscured,” or as sometimes described, a “dark,” “outlaw” zone of activity in that there remain gaps in vessel tracking data, especially in the high seas, which complicates the full assessment of IUU. ⁵ This “darkness” has economic impacts. Fisheries economists have estimated as much as $83 billion dollars are lost annually because of IUU fishing. ⁶ However, tremendous economic value continues to be captured by the sector (406 billion USD global landed value of fish and between 225 and 240 USD billion in total economic activity). ⁷ Moreover, it has been recognized that fisheries labor is at risk as a result of systemic maritime injustices because of the inherent challenges of work and the extent of human exploitation at sea. ^{8 , 9} The Global Slavery Index estimates 40 million people are experiencing abusive labor practices across the textile, agriculture, construction, and fisheries sectors. ¹⁰ Moreover, reports continue to reveal egregious human abuses at sea, such as unsafe and exploitative working conditions, but also poaching and misreporting of catch documentation, and as a result, evidence continues to emerge that shows the ocean complicates the implementation of infrastructures needed for protecting workers across fisheries. ¹¹ Such patterns have led sector practitioners, activists, and policymakers to seek remedies for maritime social problems, such as the proposals for electronic monitoring (EM) to be deployed for greater catch documentation, digital auditing, and most recently, to prevent labor abuses. ¹²

This policy brief elaborates on EM as recent trends show that this remedy is quickly gaining traction in the attempt to eliminate maritime “fishy business,” ¹³ that is, the systemic maritime-environmental injustices to labor combined with the overexploitation of fish stocks. Broadly defined, EM refers to television systems, digital cameras, global positioning systems, data storage, and transmission hardware, as well as analytical software installed on boats to surveil vessel activities such that “the overwhelming focus of catch documentation, which focuses on the tracking and tracing of fish or viable seafood products from point of capture to unloading for market sale, in recent years has been on the development and application of cameras and associated EM techniques.” ¹⁴ However, what is notable and what this paper draws attention to, is how trials are now being launched to not only watch fish but watch people. This has brought to attention the notion of “human bycatch,” or the capture of human activities on vessels via the use of EM devices. ¹⁵ Proponents argue this turn to EM may promote principles of human rights, environmental justice, and greater transparency by delivering equitable representation in maritime labor processes which would entail encouraging decent work in fisheries, expanding social protections for workers, and create the potential for increased dialogue between parties at different levels of society and workers. ¹⁶ Some policy principles already exist that could be brought to bear on this situation, such as the International Labor Organization’s decent work agenda, ¹⁷ or nongovernmental initiatives, such as the Monterey Framework for Social Responsibility. ¹⁸ Yet, as the brief explores, no frameworks for how EM could be justly applied exist and the evidence for electronic monitoring’s possibilities remains fragmented. Specifically, future work will need to carefully consider on-board dynamics, and the implications of how data are produced and managed.

FINDINGS

Species verification and identification are the most common uses of EM to date. In this way, EM was designed to illuminate bycatch—the unintended catch of nontarget fish and wildlife that are discarded by fisheries—in extent and kind (e.g., mammals and seabirds). ¹⁹ Therefore, the ability to utilize digital means to monitor catch, either in areas within national jurisdictions (e.g., exclusive economic zones, which are areas of the sea in which a sovereign state has exclusive rights regarding the exploration and use of marine resources) or areas beyond national jurisdictions (ABNJ, regions of the ocean for which no nation has sole responsibility for management), shows there is tremendous potential to “supply data required for fundamental scientific, compliance monitoring and ecological and social sustainability assessment applications.” ²⁰ EM for catch documentation and digital auditing is “being used to complement conventional human onboard observer programs to initiate at-sea monitoring where none previously existed.” ^{21 , 22} Within this context, it is worth noting that reports continue to surface about the dangers in industrial fisheries, not only for workers but also for observers. ²³ These dangers include the fact that owners and captains of industrial fishing vessels have, historically, been resistant to the costs and logistics of on-board observer surveillance of their activities. ²⁴ Tensions with on-board observers have resulted in cases of harassment, crime, and physical violence, including assault and murder. ²⁵ As a result, some scholars and practitioners think that EM could increase observer safety at sea or eliminate this labor cost altogether by virtue of having this additional means of visualizing on-board activities. ²⁶

Several cases of EM use and trends have emerged in the past decade that are of note. For example, EM initiatives have been deployed in the tuna fishing industry, corresponding with purse seine, longline, bait boat, and trawlers. ²⁷ Early work in this sector considered the possibilities of video monitoring crew activities via on-board cameras and associated computer hardware, but the interest in human monitoring grew out of the initial concern for fish documentation via catch stations on the vessels. ²⁸ Some of these early trials and programs were managed by the scientific institute AZTI (https://www.azti.es/) and Datafish Technology Solutions (http://datafishts.com/). These did not rely on one approach, instead opting for both EM and human observation to collect data to comparatively test viability. From these case studies in the early 2010s, there was evidence of potential for EM to monitor fishing labor, total catch, and large bycatch. Limitations remained with a collection of biological samples for bycatch identification as well as the monitoring of labor activities because of fisher mistrust of these devices based on concerns about the invasion of privacy and data protection.²⁵ In the early studies, data were stored on removable hard drives which were replaced every few months, and the data were then be reviewed by private firms. In addition, dedicated software programs were developed for analysis of data and imagery.

More recently, artificial intelligence (AI) has been developed as a technological refinement of these earlier processes. For instance, FathomNet is an open-source database using image data to train algorithms to identify species from trash and other debris. ²⁹ It is thought that AI tracking programs coupled with EM may reduce incentives for illegal and unreported fishing based on an underlying theory of deterrence. ³⁰ The idea behind this is that by expanding the use of these systems and making their use known, potential criminals will be less willing to engage in risky behavior. ³¹ Of course, some argue the intervention of AI represents a “revolutionary” opportunity for conservation and sustainability, ³² commentary which is, at least in part likely because of the “hype” surrounding AI in this second decade of the 21st century. ³³

Besides the relevance of AI coupled with EM to deter poaching, on-board criminality, and IUU concerns, this has recently led to sustained interest in human tracking, which would permit recording labor issues such as on-deck violence and personal protective equipment use. This is still largely conceptual in the ocean sector, although widely discussed and already used in some terrestrial contexts like construction sites, grocery stores, and other commercial applications. ³⁴ However, a discussion of the justice and equity implications in such discourse does not (yet) feature strongly. Early case studies highlighted the marine management potentials of these technologies but did not develop a plan for managing concerns about human rights abuses and injustices for fisheries laborers. For instance, throughout the 2010s, EM applications refined contributions to market initiatives to produce seafood “sustainably,” “cost effectively,” and “transparently” for “fully documented fisheries,” which eventually led some commentators to see the potential for capturing images of people on boats as a form of “human bycatch” monitoring that could solve injustices on vessels. ³⁵ Therefore, a key question moving forward is: in whom and in what are we trusting?

Although technological forms gain and evolve in meaning through the contours of their social interactions across groups in society, technologies often manifest unintended consequences and are often rooted in systems of pre-existing social inequities. Therefore, the available literature indicates caution should be taken in technology adoption processes. For example, in their analysis of EM for Scottish cod catch quotas, fisheries scientists show that fishers and captains are already time and resource strapped. ³⁶ Potential financial burdens of installing EM include camera maintenance, data transmission costs, labor associated with data review, and software licenses, although it is more cost effective than on-board observation programs in the mid-to-long term. ³⁷ In addition, widespread EM implementation will require labor to properly adjust cameras, make sure they are unobstructed, and generally ensure the quality of data collected. Fisheries policy scholar Nathalie Steins and colleagues concur, revealing in an interview study that “being able” cannot be understood as “willingness” when it comes to the introduction of monitoring technologies; fishers explained that they understood such implementations as “top down” reactionary policies to regulatory and market pressures toward an ill-defined notion of “transparent” fishery practices. ³⁸

Such findings cohere with attitudinal studies of EM, where the deployment issue has manifested as a mistrust of data use. ³⁹ To deploy EM, there must be reassurance that the data collected will be of importance to the fishers and will not be exploited, but rather serve to tailor reforms reflecting cooperation, as well as distributionally and procedurally just measures, or measures where all fisher participants will be treated fair and equitably with tangible benefits to the fishers. Benefits could include the decreased cost of paying observers, fisheries quotas that would incentivize higher quality monitoring, increased pay to vessel laborers for the use of EM, and assessing the documentation of labor abuses. ⁴⁰ This could also mean demonstration of the sustainability of the fishery, quota uplifts, and the ability to ensure the maximization of catch revenue. ⁴¹

There is also a question of experience. Studies have found that using EM technologies, even in trial runs, can increase fisher acceptance. This suggests attitudes of suspicion and negativity can be overcome with time. ⁴² Indeed, the need for trust across the social groups involved can be highlighted by any number of examples of the historical introductions of technologies. ⁴³ For example, sociologist Tim Clark’s work has highlighted how the Atlantic menhaden fishery, one of the oldest in the United States, underwent a process of industrialization and technological innovation that was especially punishing to black maritime labor. Clark advises to “be wary of market utopianism, technological innovation is not inherently progressive, and systemic exploitation still matters” (648). ⁴⁴ Importantly, not only are Clark’s insights about a certain optimism that sometimes surrounds technological innovation but Clark’s work is in line with much in critical social science that points to issues of public and private ownership of technology in environmental sectors. ⁴⁵ So far, as ocean governance scholar Lauren Drakopulos and colleagues have shown, ocean data science initiatives, of which EM is one, are almost exclusively led by the private sector and by powerful Global North actors. ^{46 , 47} Although such initiatives can be well-meaning, their underlying profit motivations may not place the needs of workers and the marginalized at the center of any initiative, such as EM. More discussions on how the public and private sector can interact in an effective manner will be needed in the future as this implies how workers on vessels are treated and how data are stored, processed, and distributed.

With regard to the questions of risks and benefits of EM, the existing studies indicate that trust will remain an important topic to consider. Trust is multidimensional, whereby a common understanding is needed that the economic, technological, and environmental goods of the sector will be shared responsibly and equitably. For example, fisheries data manager Sara Orofino and colleagues emphasize that trust at multiple levels of governance are essential to EM. ⁴⁸ While it is not uncommon to find discussions of maritime social relations corresponding to classical patrimonial logics, with understanding reached by “a gentleman’s handshake” where trust is high from frequent and consistent contact between individuals, ⁴⁹ it remains that interpersonal cooperation is the surest way to innovation and problem solving, indicating that it is at this level of direct fisher engagement that future technological forms can be implemented. ⁵⁰ For instance, in British Columbia, Canada, for crab fisheries, the key to “the success” of their EM initiatives was a “progressive discipline structure” that was focused on the pillars of fisher trust and “cooperation” and the review of human data by fisheries agencies only for “specific compliance events” or “mutually agreed upon incidents” whereby “concepts would could be applied more broadly to not just fisheries such as the transatlantic tuna fleets.” ⁵¹

STAKEHOLDERS

A plethora of stakeholders must be considered, including private firms producing EM technology, the captain, crew, and fisheries observers on-board vessels, civil society organizations that partner to trial EM, data reviewers, legal advocates, and port-based remediation organizations that help fishers bear their rights, research institutions, and governmental authorities or other fishery management organizations. Across this suite of actors, understanding how they relate to procedural possibilities to address abuse in the workplace will be essential as port-based connections to local legal experts will be crucial given that potential abuses may occur in areas beyond national jurisdictions, therefore necessitating a flexible approach. The self-advocacy of workers, sometimes linked to discussions of “worker voice” in addressing injustices needs to be supported as well, but this can be done through remediation services. Such services take a variety of forms, from connections to religious organizations, labor unions, or local community groups. The key policy pathway forward will be to systematically identify which actors exist and what capacities they hold in any given trial or future implementation project. From there, the maritime policy and advocacy community can better assess how to support maritime labor. Technological implementation needs to have clear guidelines that produce dialogic processes between all parties.

POLICY RECOMMENDATIONS

If EM will increase our vision on and around the oceans, it is necessary to explore how and why we interpret and act so that we are not simply watching crises unfold. In this spirit, several policy recommendations can be outlined.

First, privacy and trust cannot be solved expeditiously, but demand the articulation of multiple levels of action. The core ethical question at the heart of EM is about the imposition of civil liberties. The right to privacy is complicated by the quasi-private nature of the ship, which is a property relation between captain and crew who may (or not) be fairly remunerated. Although EM can record, it cannot bring about justice automatically in the case of labor abuses or illegal activities, especially if data access might entail privacy infringement. The literature shows that various private firms continue to develop automated systems that can achieve algorithmic anonymization of video via “identity management technologies.” ⁵² However, while this represents an opportunity for innovation, it is expensive and the underlying reason for implementing such algorithmic video systems is ambiguous as to whether it is truly for worker safety or securitizing firms from legal action. ⁵³ In the end, such an approach may not be in line with any of the principles of the International Labor Organizations’ decent work agenda or “worker-driven” approaches to labor.

Second, the data trail of EM has multiple dimensions, including technological limitations and bias. Iteration with experienced technicians is necessary. Common technical problems of EM include lack of camera capacity (blurry images, not enough zoom to achieve granular detail for actioning grievances), storage problems (predefined time frame to deletion of data, data format incompatibilities), and lack of organizational capacity to analyze the data (enormous data flow, specific training is requirements). Decision-making with experienced computer scientists and engineers is necessary for successful implementation. Moreover, studies in other areas such as facial recognition have shown biases, such that racialized and marginalized groups are not “identified” by computer programs in the same manner as whites. This is a potential concern that has relevance for future deployment of EM and AI-based monitoring. ⁵⁴

Third, EM implementation must be discerning in its partners, specifically in the surveillance industry, so as to align with human rights and justice commitments. EM in any form is a surveillance instrument. The firms used to deploy technologies, from the cameras themselves, to the anonymization software, to data storage and protection must be scrutinized. Failure to understand the context in which firms are operating, especially to test and trial technologies before deployment in fisheries, has potentially grave ethical implications. For example, if activists, nongovernmental organizations, private firms, and governments are committed to human rights and justice, they should not want to be linked in any way to the world’s “laboratories of war” as part of global processes to securitize everyday existence. ⁵⁵ Partners should be able to demonstrate how they are not linked to such processes and technologies. As yet, it remains opaque how one could actually assess the underlying technological supply chain, as there do not yet appear to be any definitive sources of information that documents the origins of specific EM technologies and firms that are being used, or have been trialed. Therefore, this could be an active area of inquiry in the future. However, the National Oceanographic and Atmospheric Administration published a bulletin on approved firms in 2024, but the logic behind their approval or any in-depth information about the firms remains unclear, and they do not provide any basis for independent verification. ⁵⁶

Fourth, alternative routes to facilitating remedies deserve exploration, such as greater WI-FI implementation and even social media (although this involves yet another layer of private sector partnership) to facilitate communication between fisheries stakeholders, especially laborers and loved ones, as well as greater involvement in the implementation of EM. Studies have suggested this as an effective strategy for engaging discussions bridging support for new technologies in the fisheries sector. ⁵⁷ Specifically, the support of a “worker voice” through social media use has been discussed elsewhere and deserves greater exploration specifically concerning EM. ⁵⁸ The underlying idea with such initiatives is that workers should not only have more contact with the outside world and thus more autonomy but also potentially the ability to have a greater stake in the implementation of EM. This has seen some success in other forms of fishing technologies that aim to regulate bycatch and the behavior of fishers on vessels, such as in the implementation of “green” solar-powered lights on gill nets in Baja, California. ⁵⁹ These lights reduced bycatch and were accepted because of building trust with local stakeholders through the implementation process. Although this was at the small-scale-fishery level, the adoption of worker voice that uses multiple avenues (public meetings, frequent personal engagement) could help with deployment of EM surveillance on a larger scale.

Despite the potential benefits of its deployment, the use of EM at sea needs to be treated with caution and care across multiple levels of society in order to prioritize the rights, dignity, and fair opportunities for workers.

AUTHORS’ CONTRIBUTIONS

B.F.O.: Conceptualization, investigation, project administration, supervision, methodology, writing—original draft, writing—review & editing. N.K.: Conceptualization, project administration, writing—review & editing.