Six weeks out: AFMA's electronic monitoring deadline and what it means on the deck

In April, we wrote about AI in Australian aquaculture and where digital twins, automated feeding, and underwater cameras are starting to make a practical difference in the farmed-seafood sector. The other half of the fisheries story, the wild-catch side, is now on a much harder timeline. The Australian Fisheries Management Authority has been preparing operators in five Commonwealth fisheries for mandatory electronic monitoring for months. The first deadline arrives on 1 July, six weeks from now.

For operators in those fisheries, the practical reality of what changes on the deck is already clear. The AI part of the story is genuinely interesting, but it is moving more slowly than the regulatory rollout. It is worth separating the two.

What is changing, and when

The AFMA Commission has approved expanding mandatory electronic monitoring across five additional Commonwealth fisheries: the Great Australian Bight Trawl sector, the Commonwealth Trawl sector, the North West Slope sector, the Western Deepwater Trawl Fishery, and the Northern Prawn Fishery. The phased rollout runs from 1 July 2026 through 1 April 2027.

This is in addition to the fisheries where electronic monitoring is already operating: the Eastern and Western Tuna and Billfish fisheries, the Gillnet, Hook and Trap sector of the Southern and Eastern Scalefish and Shark Fishery, and the Midwater Trawl sector of the Small Pelagic Fishery. Those operators have lived with cameras on their vessels for several years now. For the five fisheries coming on next, the change is real and immediate.

What that means in practice is multiple cameras on each vessel, recording continuously while gear is in the water, with footage offloaded periodically and reviewed against logbook records. The initial focus is on verifying logbook reporting of endangered, threatened and protected species, with the longer-term plan being to expand into discarded catch estimates and broader data collection.

The federal government has committed $20 million to the expansion. AFMA has also taken a significant operational decision in 2025 to bring its electronic monitoring review work in-house, rather than continue relying on third-party reviewers. That decision tells you something about where the agency thinks this is heading. When a regulator pulls a function back in, it usually means they expect the technology underneath to become more central, not less.

Where AI sits in this

AI and machine learning are part of the AFMA program, but in a more measured way than the marketing language sometimes suggests. The agency's own position is that it integrates AI to streamline the review of extensive video footage and enhance the efficiency and accuracy of compliance evaluations, but does not use AI in compliance, auditing or decision-making processes without human oversight.

Translated into practical terms, what AI is doing today is helping human reviewers move through hours of vessel footage faster. The systems flag events that look like a haul, a discard, or an interaction with a protected species, and a human reviewer confirms what is actually happening. The reviewer remains the decision-maker. The AI is the assistant, not the auditor.

This is consistent with what we have seen in the international literature. Companies like OnDeck AI, which started in fisheries electronic monitoring, are now experimenting with models that can analyse video without species-specific training data, which would significantly reduce the cost and time of standing up review systems for new fisheries. The pace of improvement is real. The point at which an Australian fisheries officer is comfortable letting an AI make a determination without human review is not yet here, and probably will not be for some time.

For operators, this distinction matters because it shapes what the system can and cannot do to you. The cameras are recording. The AI is helping reviewers find the relevant moments. The compliance decision is still being made by a person looking at video and comparing it to your logbook.

What operators actually need to do

For an operator in one of the five expanding fisheries, the practical work falls into three categories.

The first is the hardware itself. AFMA has been running approvals and installations through accredited providers, and the responsibility for having functional equipment on the vessel sits with the operator. By 1 July, the cameras need to be installed, calibrated, and tested in any fishery where the start date applies. The deadlines for the other four fisheries roll through to April 2027, but the practical preparation work is the same.

The second is logbook discipline. The whole point of electronic monitoring, from a compliance standpoint, is to verify what the logbook says happened. If the logbook records are sloppy, inconsistent, or incomplete, the gap between what the camera shows and what the logbook says becomes a problem. Operators who have run tight, accurate logbook practice for years have very little to adjust. Operators who have been more relaxed about the paperwork side now have a strong incentive to tighten up before the cameras come on.

The third is the practical reality of vessel routine. Cameras change behaviour, both for the operator and the crew. Where on the deck protected species interactions are handled, how discards are released, the timing of gear deployment and retrieval, all of this is now being recorded. Most of the operators we have spoken with in the older EM fisheries say that within six months the cameras become invisible and the routine settles down. The first few months are an adjustment.

What this means for the broader sector

Beyond the five fisheries directly affected, there is a quieter signal here for everyone else in Australian seafood. Electronic monitoring has been creeping outward from the original tuna and billfish trials for nearly a decade. Each expansion makes the underlying infrastructure of cameras, footage handling, AI-assisted review, and in-house regulator capability more capable and more normalised. Operators in fisheries not yet covered should expect that within five years, the question will not be whether monitoring is coming to their fishery, but in what form.

There is also a useful lesson here about how AI is actually entering regulated industries in Australia. The pattern is not a sudden replacement of human judgment with machine judgment. It is a slow, careful augmentation, with the human reviewer firmly in charge and the AI doing the volume work. This is what mature AI deployment in a regulated sector tends to look like. It is less dramatic than the headlines suggest. It is also more durable.

What to watch over the next twelve months

Three things are worth keeping an eye on between now and mid-2027. The first is how quickly AFMA's in-house review capability scales up once the cameras are running on the new fleets. The volume of footage to be processed is substantial, and the speed at which AI-assisted review can handle that volume will shape both compliance outcomes and operator experience.

The second is what happens with discarded catch data. Once the EM program moves past its initial focus on protected species and starts producing structured data on discards, the implications for stock assessments and quota management become significant. That is where the real long-term value of the program sits, and it is worth understanding before it arrives.

The third is whether the underlying technology stack gets exported. Australia has been quietly building a fisheries monitoring capability that is internationally credible. If AFMA's in-house tooling proves itself, there is a plausible path to other jurisdictions licensing or partnering on the platform. That is not certain, but it is now a more realistic possibility than it was twelve months ago.

For operators in the five expanding fisheries, the immediate work is much simpler. Have the equipment ready, run the logbook properly, and treat the cameras as a permanent feature of how the work gets done.