Self-exclusion only protects people if it is enforced. Here is how facial recognition matches a face against a self-exclusion list at the door — and why an on-premise design keeps the program privacy-first.
Self-exclusion programs let people who are worried about their gambling ask a venue to refuse them entry or service. It is one of the most important harm-minimisation tools in Australian gaming, and it depends entirely on enforcement. A signed exclusion form is only as effective as the venue's ability to recognise that person the next time they walk through the door.
In practice, that recognition has always been the weak point. A list of names and printed photographs in a folder behind the bar cannot keep pace with hundreds of patrons arriving on a busy Friday night. Floor staff change shifts, faces blur, and excluded patrons slip through — not because anyone is careless, but because manual identification simply does not scale. Facial recognition technology (FRT) is increasingly used to close that gap.
Self-exclusion sits within a broader harm-minimisation framework. A patron — or in some schemes a family member, through a third-party exclusion — formally requests exclusion from one or more venues for a set period, and the venue then takes reasonable steps to detect and turn away that person. The intent is compassionate: the system exists to honour a request for help, not to punish anyone. The challenge is consistency, because the commitment applies across every trading hour, at every entry, regardless of how busy the venue is or who is rostered on. That is exactly the kind of repetitive, high-stakes vigilance that automation supports well, provided it is deployed responsibly and with the right privacy safeguards.
The mechanics are more constrained than most people assume. A smart camera at the entry captures a face. The system converts that face into a mathematical representation, sometimes called a template or embedding, and compares it only against the venue's enrolled self-exclusion and high-risk lists. If there is a confident match, staff receive a discreet alert so they can step in early and manage the interaction with care.
The technology should make a compassionate intervention possible — quietly, before a person reaches a machine — not turn the door into a checkpoint.
This is the question that matters most, and the answer is the heart of a privacy-first design. The only biometric templates stored are those of people who are enrolled on the register because they are enrolled on the register. Everyone else who walks past the camera is compared and immediately discarded. Patrons who are not enrolled are never identified, never tracked and never recorded. The system is not a surveillance dragnet, and it is never used for marketing or loyalty profiling.
Because Ottica processes everything on a local server inside the venue, those templates and feeds never leave the premises for a cloud service. The data that the venue is accountable for stays where the venue can govern it, which materially reduces breach exposure and keeps the data under the venue's control.
A self-exclusion program is also only as good as the list it checks against, which is why database synchronisation matters as much as the matching itself. Ottica synchronises with a venue's existing self-exclusion register in real time, so a new exclusion takes effect automatically — without staff re-importing spreadsheets or copying photographs — and when a person's exclusion period ends, the same governance removes them. Done well, facial recognition for self-exclusion is not about watching everyone; it is about reliably honouring a small number of people who have asked for help, while leaving every other patron completely anonymous, every hour the doors are open.
A single accuracy percentage tells you almost nothing. Here is how to read independent face-recognition evaluations — verification versus identification, FMR and FNMR, and demographic differentials — and what to actually ask a vendor.
Read insightWhere a face is processed shapes your latency, your breach surface, your data residency and your privacy posture. An honest look at the trade-offs — and why Ottica runs everything on a local server.
Read insightBook a private demo and we will show you Ottica AI working on real venue scenarios — entirely on-premise, with nothing leaving your site.