AI drowning alerts are firing in Australian pools — here's what the system can't do for you

Two Australians are alive because an underwater camera noticed something a lifeguard didn't.

That's not a small thing. According to 7News Australia (https://7news.com.au/sunrise/inside-the-lifesaving-ai-technology-that-has-already-saved-two-australians-as-it-rolls-out-nationwide-c-22710536), AI-powered detection cameras are now running in roughly 100 aquatic facilities nationwide. The system watches for movement patterns that suggest a swimmer is in distress, then pushes an alert — including the pool grid location — to a guard's smartwatch within seconds. Early results are credible. Adoption is accelerating. And the conversation in aquatic safety circles has, reasonably, focused on how well staff respond when that alert comes through.

But there's a parallel question that's getting less attention: what happens in every part of the venue the camera doesn't see?

The camera's scope is narrower than most people assume

The AI detection system is built for one job: identifying in-water distress. It monitors underwater movement patterns and flags anomalies. It does not watch the pool deck. It does not cover the change rooms, the café, the car park, the entry gate, or the grandstand seating at a competition venue. It cannot detect a patron who collapses on a bench, a child who slips on wet concrete near the diving board, or an altercation developing at the turnstile.

This is not a criticism of the technology. Purpose-built tools work because they're focused. But aquatic venues are large, multi-zone environments, and safety planning that treats the underwater camera as the primary safety layer is missing most of the picture.

Royal Life Saving Australia's 2025 National Drowning Report recorded 357 drowning deaths in 2024-2025, running 27 percent above the 10-year average. Thirty-five occurred in swimming pools. At the same time, Australian lifeguards manage around 14,000 medical episodes annually across aquatic settings. Many of those episodes happen outside the water entirely.

Where the gaps tend to show up

Deck-level medical events. Cardiac episodes, heat exhaustion, and diabetic emergencies happen poolside. They don't trigger an underwater camera. They depend on a guard or staff member being in the right place, trained to recognise the signs, and knowing the facility's emergency action plan well enough to act without hesitation. If your EAP was last reviewed before AI cameras entered the picture, it was probably also last reviewed before your current attendance figures were normal.

Entry and exit pressure points. Busy public pools concentrate people at gates, kiosks, and change room entries. These are the zones where slips, crowd compression, and interpersonal incidents tend to cluster. They're also the zones farthest from a lifeguard's direct sightline when attention is drawn toward the water. Camera systems that cover entry points are a separate infrastructure question from drowning detection, and many facilities haven't connected the two.

Staffing ratios during an active rescue. When a lifeguard enters the water on a confirmed distress alert, deck coverage drops. This is a structural problem that technology doesn't resolve. A facility running minimum staffing has no automatic answer for who watches the rest of the pool, who manages panicked bystanders, and who makes the call to emergency services if the situation develops. That answer has to be built into the plan before an event, not worked out during one.

After-hours and low-traffic periods. AI detection cameras run continuously. Staff don't. Facilities that operate early morning lap sessions or late evening programs with reduced guard numbers are extending the camera's coverage window without necessarily extending the human response capacity to match it.

What good multi-zone safety planning looks like

The underwater camera is one layer. Effective aquatic venue safety stacks several layers on top of each other, each covering ground the others don't.

Visual surveillance of non-water zones — whether through staff positioning, CCTV, or both — fills the deck and entry gaps. Clear communication protocols ensure that a guard responding to an in-water alert can hand off deck responsibility to another staff member without a gap in coverage. Emergency action plans are reviewed regularly and reflect the actual technology and staffing configuration in use today, not two years ago.

Training needs to span all of it. Guards who are drilled on the AI alert sequence but have never walked through a poolside medical scenario are half-prepared. The same is true in reverse.

This is the kind of layered thinking XGuard brings to aquatic and public venue safety planning. Technology sets the floor on detection. The response structure built around it determines whether that detection translates into a good outcome every time, not just when conditions are ideal.

A practical checklist for venues currently running or adopting AI cameras

Before your next busy season, it's worth working through a few questions your current safety review may not have covered.

Does your emergency action plan reflect the AI alert workflow, and has it been updated to account for deck coverage during an active in-water response? Have you audited CCTV or staff positioning coverage of non-water zones in the last 12 months? Do staff in non-lifeguard roles — front desk, café, facility management — have a clear role in the emergency response sequence? Are your staffing ratios adequate when a camera alert pulls a guard into the water at peak capacity?

None of these questions have complicated answers. They do require someone to ask them before an incident makes them urgent.

Pro tip: Run a scenario drill that starts with an AI alert and then introduces a secondary event — a patron collapse on the deck or a crowd incident at the entry — two minutes into the response. Most EAPs are written for single-event sequences. Real emergencies often aren't. Identify where your plan breaks down before it matters.

The honest summary

The cameras work. The early evidence is real and the case for wider adoption is sound. But a facility that installs AI drowning detection and considers its safety obligations largely addressed has misread what the technology does.

It covers one zone, one type of event, and one phase of the response. Everything outside the water's edge is still a human systems problem. Getting that right is what separates a venue with good technology from a venue with good safety.

Need protection where you are? XGuard connects you with licensed, vetted security operators in minutes — for events, residences, retail, executive protection, and fire watch. Available globally.

Source: au-7news — 2026-08-11