3 billion people around the world obtain their protein primarily from seafood, while nearly 45 million rely on fishing for their livelihoods. And yet, at the frontline of both climate change and human over-exploitation, ⅔ of the world’s ocean fisheries (fishing industries) are at risk, while 85% of wild fish stocks are already over-harvested or depleted. The situation is so dire that in 2006, the journal Science warned that by 2048, there may not be enough fish left in the ocean to support commercial fishing. There are a number of efforts directed at reversing these trends, including government quotas for the number of fish species that can be caught, better monitoring of illegal fishing vessels, the creation of marine protected environments, decreasing use of the most destructive types of fishing, and the growth of fish farms, among others. All of these solutions, however, are dependent on having the right information, which, too often, is missing. Karen Sack, the Managing Director of advocacy organization Ocean Unite, notes, “It’s amazing that we can track a leaf of spinach from farm to table, but often cannot unscramble the supply chain for fish.” This is highly problematic for fisheries management, as well as ocean conservation efforts, which both require accurate, timely, and usable data about many aspects of the fish supply chain from the global fish population, the origins of caught fish, or even whether the fillet that ends up in a grocery store aisle is what it claims to be.
According to Tim Fitzgerald, the Impact Director at the Environmental Defense Fund’s Fishery Solutions Center, “Traditionally, catch data — the points where fish were caught, what species [were caught], average fish size, and health of the population — were obtained by the fish coming out of the ocean and [fishermen] counting them.” But self-reported information, known in the industry as “fishery dependent data” can only provide a partial picture of the fish population. Additional methods of surveying exist, such as catch and release programs as well as sonar, but these too also provide partial data and, in the case of sonar, is expensive and only effective with schools of larger fish. “In the next ten years,” says Fitzgerald, “innovation and growth will be in methods that don’t require fish to come out of the water.”
And this is where technology and, specifically, coding can play a role.
“Coding for Fish”
The State Department’s annual Fishackathon challenge drew over 1000 coders from 43 cities in 2016. The aim was to facilitate the prototyping of workable apps and gadgets to solve the information-deficit problem that exists at every level of the fish supply chain. The solutions included the “Mobile Fish Management System”, which classified, geo-tagged, and provided information on catch data for any fish, based simply on a photo taken on a fisher’s mobile phone; “Tap-a-Boat”, a mobile game that outsources the monitoring for illegal fishing vessels to players on the Internet by encouraging them to tap suspicious-looking ships in satellite imagery which, after being processed using statistical tools, would then be passed on the appropriate authorities for investigation; and “Dory”, an app that trained IBM Watson’s visual recognition technology to distinguish between different fish fillets using photos submitted by users.
In describing the rationale behind their solution, Alexi Surtees, part of the team that created “Dory”, sums up technology’s role in improving transparency in the supply chain: “Even a well-trained human eye cannot reliably tell the difference between fish (…) IBM Watson is better at deciphering [and] was chosen for accuracy. Its overtime machine learning would improve our database, enabling us to incorporate other data such as geography, sales, price, and so forth.” In other words, one of coding’s advantages is in the increased processing power of algorithms, artificial intelligence, large servers, and crowd-sourced information have over human effort alone. Fish conservation becomes a coordinated effort of a hive mind, as opposed to siloed efforts with disparate data points.
But the result of the code — the mobile apps or gadgets — are wielded by an end user and so, also need to be easy to use. Surtees continues, “From a user experience perspective it is easy and quick to snap a photo. Rather than provide a step by step [fish] identification guide, the technology does the identifying…”
Coding is Not the Panacea
The Fishackathon is joined by a number of other competitions encouraging tech-based solutions for fish, including ones run by XPrize and The Nature Conservancy. But while coding can play a key role in capturing, sorting, and analyzing information, it is not the panacea.
Sack says that “a combination of technology and policy measures are key.” After all, not everyone will be eager to provide information, especially if he or she has something to hide –often the case with ‘illegal, unreported, and unregulated’ (IUU) fishing. According to Sack, mandated regulations can help by requiring “unique vessel identifiers – numbers that cannot be changed” that, used in conjunction “with electronic log books so that data on what is being caught, where and by whom can be transmitted in real time to potential buyers.”
In other words, coding can design programs to make sense of the data, but it is (partially) up to governments to mandate, and consumers to demand the data.
Fitzgerald agrees; technology provides information that will “make our [the conservationists’] job easier” but that information “is still going to require face-to-face and interpersonal work. A lot of decisions are political.”
And with the number of stakeholders in the oceans, from commercial fishermen, environmental groups, governments, labor organizations, to consumers, there are politics at every level.
But Fitzgerald is optimistic. “Fish populations can recover in less than ten years.”