IBM has patented a new tool to help detect fraudulent activity and improve security within the cloud.
IBM's new analytical service is aimed at website operators, cloud service providers and app developers, and looks to detect unusual behaviour online before triggering a secondary authentication measure. The follow-up authentication would take the form of a security question relevant to the real user.
In practice, the tool should protect against fraud within a network, while having no impact on an authentic user.
"When individuals access banking or shopping site, they subconsciously establish characteristics of how they interact with the site, such as clicking certain areas more often than others," said an IBM spokesperson.
The company claimed that during development, the software had little effect on genuine users, while successfully identifying fraudulent activity.
Big Blue's Keith Walker, the co-inventor of the patent added, "Our experience developing and testing a prototype, which flawlessly confirmed identities, shows that such a change would more likely be due to fraud.
"For example, if an individual suddenly changes how they interact with an online bank or store, such as due to a broken hand or using a tablet instead of a desktop computer, I want these websites to detect the change, and then ask for extra identity confirmation."
It is not yet known when the invention will be generally available, but this is one of 290 patents that IBM has developed to deal with fraudulent transactions, as the company continues to invest heavily in security research.