Faculty, the artificial intelligence firm that previously worked for Dominic Cummings’s Vote Leave campaign, has been awarded a public contract to help the government assess the use of machine learning to secure Britain’s post-Brexit fishing waters.

According to the description of a £50,000 contract awarded to Faculty in March 2020, the UK’s Department for Environment, Food and Rural Affairs (Defra) was seeking to better understand how artificial intelligence systems could be leveraged to spot vessels engaging in illegal, unlicensed or unreported fishing in Britain’s waters.

The contract itself, published on the government’s Contract Finder platform, established that Faculty would “assess the potential of innovative technological solutions (specifically, machine learning) for sea fisheries control and enforcement”. The company was listed as a processor of public data and the data of people involved in sea fishing during the course of its work.

Fish has emerged as a particularly slippery issue and threatens to sink the prospect of a post-Brexit trade deal with the European Union. On Friday, in a speech declaring that the UK should ready itself for “an Australia-style” future with no trade deal, prime minister Boris Johnson listed taking back control of fisheries as one of the key positives of crashing out of the EU’s single market. If no deal materialises by the end of December, Britain will have to extricate itself from the EU common fisheries policy after over four decades, with nothing to replace it. The UK is already poised to start patrolling its seas to keep out European vessels, which will no longer be automatically allowed to fish in British waters under a no-deal scenario.

Scott Edwards, an expert in maritime security from the University of Bristol says that AI is increasingly regarded as a promising tool to make sense of vast reams of maritime data pool coming from satellites, transponders, aerial photography and various other sources. “Machine learning [can be used] to identify patterns and spot which vessels stand out as potentially being involved in illegal fishing,” Edwards says. Algorithms using machine learning might be able to quickly single out vessels which are moving evasively or appear to be tampering with their Automatic Identification Systems, and flag them as suspicious.

Faculty’s maritime interests don’t stop with the Defra contract. Earlier this week, the firm announced that it has started working with the Maritime and Coastguard Agency to provide AI tools to boost “search and rescue” operations – an increasingly hot-button topic in the wake of the arrival of over 4,000 asylum seekers in small boats since the start of the year. Over the past few months, Faculty has been awarded some lucrative government contracts to help officials deal with the coronavirus crisis. Reports have highlighted the firm’s close links to senior Number 10 advisor Cummings, who hired Faculty when heading the pro-Brexit campaign in 2016, and Ben Warner, Cummings’s right-hand man – and brother of Faculty’s CEO Marc. Faculty did not respond to a request for comment by the time of publication.

Machine learning wasn’t the only technology Defra was keen to delve into when charting the UK’s post-Brexit fishery future. Also in March, a £39,000 contract was awarded to technology consultancy Frazer-Nash, aimed at helping Defra make sense of how drones could be used to better defend Britain’s fish stock. William Barnes, Frazer-Nash’s Unmanned Aerial Systems service lead, who was part of the team working on the Defra project, explains that the usage of drones in fisheries enforcement is evolving and hard to navigate.

“There is a breadth of different technologies – that’s really the question that Defra asked: they recognise the relative difference between technologies from different aircraft, different sensing platforms, different ground stations,” Barnes says. “And how you put them together depends on what is needed. It’s never going to be a one-size-fits-all right.” For instance, drones could be used in a rapid-response way – alerting the navy to fishing boats illegally hauling fish and scallops in – or as a planning tool, keeping an eye on general trends taking place in the British seas.

While several countries have been looking into the use of drones to surveil their fishing zones, the technology is still relatively untested, says Edwards. “There have been a couple of countries using camera-equipped drones for this: the Seychelles is probably the flagship one with a project called Fishguard,” he says. “It still hasn’t been fully implemented, and they had to do a lot of testing because of problems like bad weather. It’s not actually that easy to establish.”

Edwards notes that drones and AI have to go “hand in hand,” with images captured by drones then used to refine machine learning algorithm’s detection mechanisms, and basic AI helping drones on a patrol to decide which vessels to track.

A Defra spokesperson says that the department, alongside the Marine Management Organisation, a public body that helps regulate Britain’s waters, is “exploring how technology can innovate and enhance fisheries monitoring and surveillance methods.” This includes sensor imagery, autonomous monitoring and machine learning, the spokesperson adds, which could potentially complement existing systems.” More details on this would be available later this year, the spokesperson says. Two more Defra contracts on Contract Finder, whose descriptions are almost identical to those of the contracts awarded to Faculty and Frazer-Nash and appear as “closed” – a term suggesting that the contract hasn’t been awarded.

Edwards says that while the potential of drones and AI for fisheries enforcement is promising, full implementation will take time. In the meantime, the UK will likely focus on beefing up patrols. Currently, the navy’s Fishery Protection Squadron counts five ships, although the government has announced plans to treble that capability. “The drones can help but only if they've been implemented as part of a wider approach,” says Edwards.

Gian Volpicelli is WIRED's politics editor. He tweets from @Gmvolpi

This article was originally published by WIRED UK