Commonwealth Bank of Australia is implementing artificial intelligence technology and machine learning techniques to detect serial abusive behaviour in transaction descriptions within the CommBank App and Netbank.
The Australian bank last year acted to ban customers that used transaction description fields to send abusive messages, including threates to domestic abuse victims.
A study by the bank, prompted by disturbing messages in the account of a customer experiencing domestic and family violence, found that more than 8000 customers had received multiple low-value deposits, often of less than $1, with potentially abusive messages.
The new model was developed in the CBA AI Labs, and for the first time allows the bank to proactively identify serious instances of abuse. The new tool works in tandem with an automatic block filter that was implemented last year across its digital banking channels to stop transaction descriptions that include threatening language.
Justin Tsuei, general manager community and customer vulnerability, says: “The new model, which uses advanced AI and machine learning techniques, allows us to provide a more targeted and proactive response than ever before. With this new model in place, not only are we able to proactively detect possible instances of abuse in transaction descriptions, but we can do so at an incredible scale."
Over a three month period, from 1 May to 31 July 2021, over 100,000 transactions were blocked by the bank's automatic language filter. Of those instances, the new AI tool detected 229 unique senders of potentially serious abuse, which were then manually reviewed to determine severity and the appropriate action required from the bank.
This can include setting up new safe accounts for victims and terminating banking relationships with perpetrators.
CommBank is not alone in clmaping down on technology-facilitated abuse. Westpac in Februrary introduced a new tool enabling customers to click a report button within its online and mobile banking platforms to have offensive messaging flagged and sent to a dedicated team for review and action, which may include notifying the sender's financial institution of the interaction.
Westpac is also using technology to monitor the language used in outbound payments, blocking certain transactions containing inappropriate or offensive language in real-time.