A popular subject of discussion in the maritime sector for some time now, artificial intelligence is gradually establishing itself in fish farming. Sunderland Marine Aquaculture Manager Duncan Perrin and risk management surveyors James Simison and Allan Lyons outline the applications and advantages of AI in the fish farming industry – but also offer a word of caution.
Being a much smaller, younger and lower-profile industry than shipping, aquaculture is at a less advanced stage in its adoption of artificial intelligence and related technologies. Yet as AI continues to mature and its price steadily falls, more and more fish farmers will see the advantages it offers as simply too good to ignore.AI is already being deployed in aquaculture to optimise feeding. The method sees cameras programmed to distinguish feed pellets from seaweed and other detritus installed at depth within pens to monitor feed uptake. If too many pellets pass the camera – meaning they have not been eaten – the system automatically shuts down and alerts the technician that food is being wasted. This spares staff countless hours watching screens and allows them to have several pens fed at the same time. It also saves money on feed, one of the major costs in running a fish farm.
Alongside feeding, the most crucial factor in rearing healthy fish is the quality of the water in which they are kept. Water-quality monitoring is therefore another common application for AI in aquaculture, with systems able to alert operators when, for example, dissolved oxygen levels drop below the normal range. In addition, the technology can monitor the status of the machinery that regulates water quality, notifying the technician of any potentially dangerous issues – such as a valve being left open or closed for too long or a lever being in the wrong position.
Also essential to the health of fish is delousing. In this context, AI is used to monitor the presence of sea lice on salmon, with cameras identifying the individual lice, counting and measuring them and determining their stage of development. This allows farmers to assess the risk posed by the parasites and respond appropriately. Yet some farms are now automating the delousing process, installing technology that pinpoints lice and exterminates them with lasers that, unlike certain medicines, cause no harm to the fish or the environment.
Ultimately, automation will be ubiquitous in aquaculture. Having systems monitor various processes and conditions and share information with operators is a time- and cost-effective use of AI, but programming systems to control these conditions themselves is even more valuable. Accordingly, the next step for AI in onshore farms will see the technology not only monitoring but also regulating concentrations of dissolved gases in the water. It will achieve this by activating and deactivating equipment such as aerators to maintain a healthy environment for fish.
While the above examples demonstrate the undeniable value of AI in aquaculture, the industry must take care not to become overly reliant on the technology at the expense of a skilled and experienced workforce. From an insurance perspective, leaving a small team in charge of millions of fish is simply too risky – and yet it may seem like an attractive proposition to some fish farmers.
A high level of automation is particularly appealing in offshore aquaculture, where it is costly and inconvenient to send personnel to remote locations to handle routine but labour-intensive tasks – like feeding and water-quality monitoring – that a computer can perform automatically. However, a highly automated set-up increases the threat of cyber-attacks, and in a remote setting, the physical risks of having a small number of staff on site are exacerbated.
A less tangible but equally critical issue is that of human intuition. Farmers of any kind possess an understanding of their animals that machines cannot – and might never – replicate. This instinct is passed from generation to generation and honed with experience, but an over-reliance on technology threatens to suppress it. The bulk of the responsibility in aquaculture should therefore remain with the farmers.
Nevertheless, when deployed as a tool and not as a replacement for human expertise, AI is a remarkably useful technology that promises to enhance the efficiency and profitability of the aquaculture industry as well as the welfare of the fish on which it relies.
Post written by:
Duncan Perrin
For more information, visit: https://sunderlandmarine.com