A new cyber security virtual assistant using artificial intelligence has been developed to help detect threats and guide users and organizations through the necessary steps to respond to attacks. Australian researchers from the University of Technology Sydney (UTS) led the project. The project is described as a new type of tool that can integrate information from multiple sources and provide context-aware recommendations.
AI for cyber defense
The system is called CASPER AI, which stands for Context-Aware Security Policy Enforcement and Response. Bring together your organization’s information, from identity management platforms and HR records to travel, expense data, and network logs. By correlating data from these sources, CASPER AI aims to identify anomalies that may indicate the presence of insider threats, compromised user accounts, or external compromises.
The assistant relies on large-scale language models and multimodal reasoning to analyze various datasets and determine whether a cybersecurity threat exists. According to the developers, CASPER AI can identify abnormal operations, such as the creation of fake wireless access points in public areas or suspicious behavior in temporary mobile environments, and guide users to take appropriate and compliant security measures.
“The demonstration showed how CASPER AI can detect anomalous activity within enterprise operations, identify bogus wireless access points in public spaces and temporary mobile deployments, and help teams quickly take policy-compliant security measures,” said Professor Ren Ping Liu, University of Technology Sydney project leader.
Assistance from non-experts
The development team states that one of the key benefits of CASPER AI is its understanding of context. Traditional rules-based cybersecurity tools are reactive and often miss subtle threats that bridge multiple information sources. CASPER AI aims to provide clear guidance to staff and reduce reliance on users when interpreting technical security policies in high-pressure situations.
AI assistants are built to work in a variety of environments and can reduce manual workload for security teams. The easy-to-access interface and plain language output are designed to support people with limited digital experience.
“The system reduces the need for staff to interpret complex policies under pressure by presenting recommended actions in clear, easy-to-access language,” said Liu.
Fraud and fraud detection
Although CASPER AI is aimed at implementation within organizations, project leaders also highlight the potential benefits for the individual population, particularly those at risk of online fraud and those with low digital literacy. The system is capable of detecting unusual activity related to personal accounts and can provide recommendations in a format that even non-expert users can understand. Designed to help individuals verify the authenticity of emails, texts, and calls and detect identity abuse.
“CASPER AI as a virtual assistant helps people with low levels of digital literacy and technical experience avoid falling prey to cybercrimes such as phishing and ‘smishing’ (SMS-based) fraud,” said Liu.
The Assistant explains in simple terms whether a suspicious message may be a scam and advises users if their login credentials or personal information may have been compromised. For government or other sensitive accounts, you can flag sign-in attempts from new or unauthorized locations.
Network initiatives
CASPER AI is part of a series of projects funded through the Connectivity Innovation Network. The initiative is supported by the New South Wales Government and led by UTS and the University of Sydney. Its goal is to address the challenges of digital connectivity and cyber resilience as the volume and complexity of online threats increases as communities, citizens, and government agencies become more dependent on digital services.
“It’s designed to keep government accounts safe by detecting anomalous login attempts and activity, protect people’s identities by linking information across services to flag potential abuse, and stop fraud by giving people an easy way to verify whether a message, email, or phone call is genuine before they respond,” Liu said.
