Epistemic status: I researched this in a day and a half, wrote it in a couple hours, and shared it with no one. Believe nothing and fact check everything.
Purpose: 1. This is an experiment in how quickly I can learn and write about a new cause area. Feedback would be very helpful. 2. Maybe for someone equally new to the topic of fish, it will speed up your learning!
(A summary of the documentary The End of the Line.)
Indigenous peoples subsistence fished for millennia. When European settlers arrived in North America in the late 1400s, they were stunned by the fish population. For the next four centuries, Europeans would immigrate to North America to support themselves as fishermen. In Newfoundland, cod fishing was extremely popular. 
In 1951, factory fishing of cod began.  But by 1992, cod were practically fished out of existence in Atlantic Canada. A moratorium on cod fishing began in 1994. 40,000 people lost their jobs, and protests ensued. Cod continue to be endangered today, and the moratorium has lasted over 30 years. 
In the late 20th century, data showed populations of many species of fish around the world decreasing, but the "worldwide catch" was increasing. Tracing the discrepancy to China, researchers realized that the Communist Party had been falsifying fishing data. In 2002, researchers realized that the true worldwide catch had been decreasing since 1988. Fishers panicked. 
According to Boris Worm, populations of the fish we eat have declined 90% around the world. By 2003, a third of species we eat had collapsed. All may collapse by the mid-century. 
Trawling nevertheless continues. In Senegal and elsewhere, while local, impoverished fishers lose catch every year, their governments sell fishing rights to richer countries who use supertrawlers. Immigrants who can no longer afford to live as fishers in their home countries are turned away by the same governments that sponsored the destruction of their ecosystems and jobs. 
The largest trawling nets fit thirteen 747 jets. Trawling nets plough the ocean floor multiple times a year, disturbing delicate ecosystems like coral reefs. 
Now that the populations of many predator fish, such as cod, have collapsed, smaller species of fish and arthropods have exploded. Some examples in Atlantic Canada and the USA include lobsters, Chesapeake rays, shrimp, and jellyfish. 
The Marine Stewardship Council is one body that tries to prevent overfishing and improve the sustainability of the industry. Yet 93% of the ocean remains unprotected by Marine Protected Areas.  (You could just stop eating wild fish.)
Is farming a solution?
So let's stop catching wild fish! We can just grow our own, right? We can farm some salmon. Salmon eat herring. So we will fish for some herring to feed the salmon -- crap.
Today, up to three trillion fish are killed for food every year. About half are wild and half are farmed. In fish production by weight, farming overtook wild fishing in the 2010s. Fish farming is the fastest growing form of animal food production.  The number of farmed fish alive today is approximately as many humans as have ever lived. 
Does farming reduce overfishing? No. Evidence suggests that farming either doesn't impact overfishing or makes the issue worse.  Predator fish eat small fish. We farm a lot of predator fish, and we catch a lot of small fish to feed them with. For every pound of predator fish, five pounds of small fish were used to feed them.  This amounts to hundreds of billions of individual fish caught to feed farmed fish every year. 
Putting the ineffectiveness of that aside, let's take a brief interlude to learn what it's like to be a fish.
Do fish notice their exploitation? They don't have a neocortex.  However, some species of fish have demonstrated:
- long-term memory
- social structures
- problem solving abilities
- the ability to use tools 
- the ability to hunt, which requires some ability to predict the behaviour of prey 
- adaptive emotional states
- passing the mirror test 
Fish have nociceptors, which is a type of cell that carries pain signals. Fish have demonstrated the ability to feel pain by:
- avoiding painful stimuli
- treating their injuries
- being distracted from survival behaviours by painful stimuli
- not being distracted by painful stimuli after receiving morphine 
- naturally releasing opioids 
So it's a good thing that humans always run respectful, painless farms.
What happens on a fish farm?
Just kidding. (Skip down to "Loch Duart, ikejime, & the next frontier" if you want good news.)
Fish on a farm live, and then they are killed. Let's learn about the killing first.
Fish are (usually, thank goodness) killed before we get them. The most common slaughter methods are asphyxia and live chilling. These are terrible, and take from twelve minutes to several hours. 
As fish asphyxiate, they thrash against the ice and bleed internally. 
Live chilling involves putting fish in an ice slurry in an attempt to render them insensible, or stun them, before killing them. But according to Andrés Jiménez Zorrilla from Shrimp Welfare Project, ice slurries are often used incorrectly (in the case of shrimp, anyway) and may not be effective at stunning as they are used. 
Some other slaughter methods include gassing with CO2, cutting the gills and letting the fish bleed out, and and processing alive. 
They're not always killed before we get them, though. Some people buy live fish. This means that the fish have gone through many stressful transports and will be inexpertly slaughtered without stunning. 
There are a couple "humane" ways of stunning fish. These are electrical stunning (sending a current through the brain) and percussive stunning (hitting the skull very hard).  Andrés has mentioned that implementing electrical stunning on a massive scale, such as running a current through a pond, poses significant safety challenges.  I imagine it is also difficult to accurately implement percussive stunning at scale as well.
But let's say we can efficiently stun all the fish we intend to kill. How do we go about doing that? Unfortunately, there are not clear standards for what electrical or percussive force is needed to stun each species of fish. 
We also usually can't tell whether a fish has been rendered insensible by stunning or just immobile. There is one practical way to check, which adds time to the slaughter process: checking each fish for the eye roll reflex. 
That wasn't super promising. Do fish live good lives on farms? Buckle up.
We breathe oxygen through our lungs. Fish process dissolved oxygen in the water through their gills. Each species of fish thrives at its own best level of dissolved oxygen. Unfortunately, we mostly don't know what that is for each species, beyond being somewhere between 3 and 9 mg/L, probably. What we do know is that whatever the ideal is, most farms probably aren't living up to it. 
When dissolved oxygen is too low, fish might get a feeling similar to how we feel when we can't get enough oxygen. Lots of people who have had pneumonia know what that's like.
Let's look at some more specific issues. Here are some things you will probably see on a salmon farm: 
- tens of thousands of salmon, and quite a few of them are dead
- aggression and injuries
- low oxygen and high ammonia
- little expression of natural behaviours
- no natural migration patterns
- old food
- sea lice that feed on skin and blood, leaving ulcers
- 50% of farms in Scotland have sea lice infestations at a time 
- starvation before transport, grading, and slaughter
It's not just bad for the farmed fish. Hundreds of thousands of farmed fish escape per year, spreading diseases and sea lice to wild fish, and threatening biodiversity. 
Also, it's not just bad for fish in general. Hundreds of seals and thousands of birds are also killed per year on salmon farms to minimize loss. 
We would really like to get rid of some of sources of suffering. But how? There are very few "high certainty" welfare areas for fish. In fact, there are none for the grass carp, and we farm billions of them per year.  And there are 362 species of fish we farm, all with different needs. 
Fish don't really have to be alive yet to have issues, either. Farmed fish are born with myriad problems.
Like chickens, some fish have been bred to be too big to maximize yield. This results in cataracts and heart defects. 
When these weird little guys escape the farms, they breed with wild fish, which makes the wild fish population worse at surviving and reproducing. This is called genetic pollution. 
Some farmers want to avoid the problem of genetic pollution, so they create triploid salmon, which cannot breed with wild salmon. However, triploid salmon seem to live worse lives. 
How bad is that?
Wow, that was terrible. But is it the absolute worst? Yes, probably.
In December 2018, Charity Entrepreneurship (CE) developed the Weighted Animal Welfare Index (WAWI).  It contains eight metrics, themselves scored on:
- accuracy of the measure as a proxy for ethical value
- cross-animal applicability
CE ranked fifteen animals on each of the eight metrics to create a welfare score. Wild fish scored 7th on welfare. Factory farmed fish in a traditional aquaculture setting scored 10th. Wild fish that humans catch scored 12th out of 15.
CE then combined the welfare score with population size, odds of feeling pain, neglectedness, and number of reasons for suffering. This created an overall ordering of which animals were highest priority to help.
As they say, there are plenty of fish in the sea. As it turns out, even though farmed birds may have worse lives as individuals, the scale of fish suffering is just enormous. Wild fish are particularly neglected, and suffer for many reasons.
Wild fish for human use, other wild fish, and factory farmed fish took the top three spots in CE's priority ranking of animals to help. Someone, please, help the fish.
Loch Duart, ikejime, & the "next frontier"
In 2016, someone tried to help the fish. Salmon farming company Loch Duart hold themselves to high standards in fish "health, diet, water quality, husbandry, handling and slaughter."  But they saw that their salmon were still nipping each other's fins, which is a sign of stress. Manager David Roadnight figured his fish might be bored, and he gave them some coloured balls and a kelp-like tarp. These props are both types of environmental enrichment. (The effectiveness of environmental enrichment overall has mixed evidence and depends on the species. )
The coloured balls distracted the fish from nipping each other. The tarp provided a hiding place, which allowed the salmon to express more of their natural behaviours. When fish could hide from the disturbance of passing shadows, they would stop showing signs of stress sooner after the disturbance passed.
The tarps resulted in better use of tank space, less nipping behaviour, and less fin damage. The tarps seemed to be a more cost-effective welfare improvement than the coloured balls. 
In 2018, Andrew Tsui, founder of the Ike Jime Federation, invited Vox to kill some fish with him. Ikejime is a slaughter-stunning method purported to be more humane than alternatives. The fish are stabbed through the brain and rendered brain-dead instantly. Then a wire is inserted into the spinal column to eliminate thrashing. Finally, the fish is bled. This eliminates the build up of blood and lactic acid in the tissues. 
In 2021, Vox announced that fish welfare is the "next frontier" in animal advocacy. 
Where should we help fish?
So if we are going to help the fish, where should we do it? By weight, 88% of farmed fish are in Asia, and half of farmed fish are in China.  But there may be more of a precedent for helping animals in India, where the second-most fish are farmed. 
Of the 111 billion fish farmed per year, about 3-15 billion of them are farmed in India. A person in India typically consumes more fish per year than land animals by the kilogram. The highest fish consumption is by men, people in urban areas, and people in the following coastal regions: 
- Kerala, on the southwest shore
- West Bengal, on the eastern point of the mainland (capital: Kolkata)
- Lakshadweep islands
- Andaman islands
- Nicobar islands
India seems like a promising place to work on animal advocacy because they have strict and extensive animal agriculture laws. However, these laws are rarely enforced, and the listed fines have not kept up with inflation. 
Animal advocacy is common, as is vegetarianism. However, 90% of animal advocacy groups in India focus on street dogs, and many of the rest focus on cows. 
India's government may be open to animal advocacy, but they do not seem particularly concerned about fish (as of 2021). The government is trying to increase aquaculture, which is unregulated. Fish farms in India have low adherence to the Aquatic Animal Health Code. 
Perhaps they could be persuaded. Here are some opportunities: 
- strong animal agriculture laws
- public support for welfare standards
- a culture of non-violence represented by ahimsa
- strong research labs that could innovate on alternative proteins
- concern about the health and environmental threats that have come with the recent increased consumption of meat
Here are some challenges: 
- lack of enforcement (and enforcement committees)
- poverty and malnutrition (in 40% of children!), making dietary change for animal welfare less of a priority
- legal issues around funding and censorship
- meat as a status symbol
- desensitization to animal suffering
- lack of interest in animal advocacy careers
- the small animal replacement problem as people avoid eating cows
So who is going to take on these challenges and fight for the fish in India?
Here comes Fish Welfare Initiative
I don't work for FWI, this is just my interpretation of their online documents!
In 2019, Charity Entrepreneurship (CE) wrote a report entitled "Improving Environmental Conditions."  The report found that simple environmental changes can improve welfare, and the most promising of these is improvements in water quality for farmed fish. CE estimated that optimizing dissolved oxygen levels would cost-effectively provide 14.75-21.5 welfare points per year.
Fish lives at that point would still be net-negative,  so it would be important not to increase the number of fish farmed.
CE noted, as had ACE before them , that more research would still be needed into optimal dissolved oxygen levels for different species of fish. They also noted that optimal dissolved oxygen depends on other environmental factors.
But CE found this intervention promising for several reasons: 
- funding and staffing were not expected to be an issue
- improving water quality for fish doesn't look "extreme" (although it may look unimportant)
- measuring water quality is affordable and easy
- aerators are affordable, but not too affordable, which might drive price up and demand down
- the intervention can be presented as good for farmers
- feedback loops would be quick
- it could bring new people into the animal advocacy movement
As a result of this report, Haven King-Nobles and Tom Billington founded Fish Welfare Initiative (FWI) through the 2019 Charity Entrepreneurship incubation program. Their mission was to reduce the suffering of fish as much as possible by identifying and piloting interventions. 
Tom made the case that fish welfare is an important, neglected, and possibly tractable cause area. He noted that the base of research on fish welfare had been growing, that aerators for increasing dissolved oxygen were affordable, and that stunning had already been successfully implemented on some farms. 
Although the Weighted Animal Welfare Index  had placed wild fish as a priority above farmed fish, Billington argued that FWI would have more impact targeting farmed fish for two reasons. Firstly, humans have more influence over the lives of farmed fish. Secondly, fish farming is a rapidly growing industry, much faster than wild catching.
FWI strove to answer these questions: 
- Which species should they focus on?
- Which interventions (welfare factors and campaigns) should they focus on?
- Which countries should they focus on?
The same year, William Bench had founded the Aquatic Life Institute (ALI).  ALI focused on highly tractable countries (in Europe), working mostly with corporations and certifiers. FWI would go on to work in countries with high scale and neglectedness (in Asia), working mostly with farms and farmers.
Over the last few years, FWI has grown from $50,000 in seed funding to $1,050,000 in total fundraising. They have also grown to 20 staff members. 
In FWI's 2021 retrospective, they noted that their key bottleneck was improving their welfare standard. At the time, it was not proving impactful enough to scale up. They also found that their expectations for standardization were unrealistic.  Nevertheless, they had already:
- Helped 214,000 fish
- Connected with farms
- Added several people to their team who were previously not a part of the animal advocacy movement
- Secured the first corporate commitment for fish in India
- Established the Alliance for Responsible Aquaculture (ARA)
- Authored several reports 
By the end of 2022, FWI had scaled up significantly.  They had:
- Helped a total of 1.14 million fish at a cost of $1.13 USD per fish
- Grown the ARA by 53%
- Created a second version of their welfare standard
- Signed two memoranda of understanding
- Become a partner of the Andhra Pradesh state government
They found they needed more rigorous experiments in the field and more corporate partnerships.
Between 2021 and 2022, FWI and the ARA had collected the following measurements on partnered farms to evaluate the success of their welfare standard: 
- dissolved oxygen
- total dissolved solids
- stocking density
- feeding amount
- disease presence
- lice presence
- behavioural indicators
- air gulping
- tail splashing
- primary productivity (a proxy for phytoplankton)
Future of FWI
On January 1, 2023, FWI posted that they were updating their welfare standard to make greater improvements in dissolved oxygen (and thus welfare). They added a standard of three days per month for carp to feed on phytoplankton instead of supplemental feed. Carp naturally feed on phytoplankton, so when they receive supplemental feed (to increase yield), the phytoplankton grow. This destabilizes the level of dissolved oxygen. 
This year, FWI will be evaluating Version 2 of their welfare standard and creating a Version 3. 
In addition to testing and validating Version 2 of the welfare standard, FWI intends to scale up work in India. They intend to help 1.5 million fish in 2023 and more than double the farms in the ARA. 
As they do this, they will be implementing a new theory of change-based strategy.  Their activities will be driven more by validating a theory of change for its long-term impacts than by helping as many fish as possible in the short-term. They will be moving away from choosing, implementing, and scaling interventions. Instead, they will be validating their new theory of change. 
In their old theory of change, government and corporations would partner with the ARA, require welfare standards, and help fish. In the new theory of change, FWI expects that farmer engagement will drive enrollment in the ARA, and farmers will implement welfare standards and help fish. 
I don't work for FWI, this is just my interpretation of their online documents!
- Fishing used to be pretty okay.
- Then we fished way too much!
- Wild fish welfare is terrible anyway, though.
- So we started farming.
- It did not help!
- Fish are sentient beings who feel pain, and they suffer a lot on farms, as they live and as they die.
- They might be the highest priority animal to help.
- But we can help!
- We can mostly help by increasing dissolved oxygen.
- Maybe we can help the most in India, where a lot of people care about animals but they also farm and eat a ton of fish.
- Fish Welfare Initiative and Aquatic Life Institute are focusing on cost-effectively helping farmed fish.
- Fish Welfare Initiative will be scaling up and testing a new theory of change this year.
- Donate to FWI!
- Donate to ALI!
Next steps for me
I still have a bunch of reports from Fish Welfare Initiative's research page to go through.
I'll need to read some primary research. I'll start with the works of Fish Welfare Initiative's Fish Welfare Specialist Consultant, Marco Cerqueira. Then I will branch out into related authors and anything cited in the welfare reports I've already read.
I'll also be taking a course called Large Marine Ecosystems: Assessment and Management. This might teach me more about wild fish, but I think it teach me some useful research methods in biology. There's also a unit on plankton, which is a key contributor to low and changing oxygen levels on fish farms.