The National Retail Association is pleased to have been joined last month by Queensland Small Business Commissioner, Dominique Lamb, and our Crime Consultant, Professor Michael Townsley, at the free online session for retail and loss prevention specialists. Attendees received updates on the latest criminology research, the current retail crime climate across small businesses, and academically forecasted retail crime trends.
A key call out for the industry is the emerging role that Artificial Intelligence (AI) is likely to play in preventing and responding to retail crime and the potential impacts of this engagement.
The following summarises Professor Townsley’s response.
What role is AI likely to play regarding retail crime?
It is too early to be definitive, but the pace of development is staggering.
- Video Surveillance and Analysis: CCTV systems require human monitoring, which can be labour-intensive and prone to errors. AI-enhanced surveillance systems can automatically detect suspicious activities and alert staff in real-time. These systems can recognise patterns, faces, and even predict potential theft based on behaviour or gestures. The applications for self-checkout are noteworthy – supervising customers to a level that would be cost-prohibitive for staff on a 1:1 basis.
- Predictive Analytics: AI can analyse vast amounts of data to predict which products are most likely to be stolen, when thefts are most likely to occur, and which locations are most at risk. Customer behaviour (dwell time, wayfinding, gait/walking style) could be quantified and tracked with a view to prioritise abnormal behaviour.
- Fraud Detection: AI can analyse transaction data to detect unusual patterns or behaviours that might indicate fraud. For instance, frequent returns, unusual purchase patterns, or the use of stolen credit card information can be flagged for further investigation.
From an academic perspective, it is yet to be seen how the technologies are implemented with existing systems and processes and what compensating activities are put in place to accommodate AI. Offenders are superb at probing for vulnerabilities and loopholes in systems, so it should be expected that they will locate and exploit any weak points.
If we consider the ways organised offenders may employ AI, several ideas are obvious
- Data Theft and Breaches: AI tools can be used by malicious actors to identify vulnerabilities in a retailer’s digital infrastructure, leading to data breaches. This can compromise customer and employee data.
- Fraudulent Transactions: AI can be used to generate fake invoices and account takeovers — see FraudGPT.
- Targeted Scams: By analysing customer data, AI can be used to craft highly personalised phishing scams or fraud schemes targeting specific customers or employees.
How might confrontational situations, like using self-serve gates to prevent exits, parallel concerns with AI surveillance in loss prevention?
While AI-powered surveillance can help in loss prevention, it can be seen as a form of surveillance capitalism. Unauthorised tracking, facial recognition without consent, or profiling based on behaviour can lead to significant privacy concerns and customer dissatisfaction.
The National Retail Association is continuing with state-based Retail Crime Committees, with the second round of bi-annual meetings currently underway. We look forward to welcoming industry and facilitating valuable information sharing sessions, ahead of the busiest season in retail.
