Digital Twins of AI Workers Create Emerging Insider Threats

Artificial Intelligence & Machine Learning,
Black Hat,
Events

Researchers Highlight Risks of AI Bots in Cybersecurity Contexts

Michael Novinson (@MichaelNovinson)
•
August 15, 2025


Matthew Canham, Executive Director, Cognitive Security Institute, and Ben Sawyer, Associate Professor, Industrial Engineering and Management Systems, University of Central Florida

As generative artificial intelligence technologies advance, they pose evolving threats to contemporary workplaces. The emergence of digital twins—hyper-realistic digital replicas of actual employees—facilitates the imitation of vocal patterns and behavioral trends, which could significantly impact cybersecurity frameworks.

See Also: Organizations Must Concentrate on Threat Behaviors, Not Just Flaws

While the increased efficiency offered by digital twins may be advantageous, they simultaneously introduce substantial insider threats. The complex interrelationships surrounding consent and data ownership raise critical questions about the extent to which companies may replicate human workers—an issue highlighted by experts Ben Sawyer and Matthew Canham.

“From a business perspective, it makes sense for companies to collect such data. If employees feel that their own data is being utilized to create a setup that could replace them, it may drive them to sabotage that data,” commented Sawyer.

In a recent video interview with Information Security Media Group recorded at Black Hat USA 2025, Sawyer and Canham explored various pertinent topics, including the sophistication of AI-driven phishing attacks and the potential for social engineering exploitation of digital twins. They also discussed the emergence of new legislation in previously unforeseen sectors.

Canham, who has extensive experience in cognitive security and human-technology integration, served as a supervisory special agent with the FBI before taking on leadership at the Cognitive Security Institute. His research concentrates on the cognitive dimensions of synthetic media and social engineering.

Sawyer, a human factors engineer with a focus on the interface between humans and machines, has collaborated with notable institutions, including the U.S. Air Force and the Massachusetts Institute of Technology. He brings his unique expertise in mathematics and research to bear on projects for various Fortune 500 companies.

Source link

Related Posts

Iranian Hacker Admits Guilt in $19 Million Robbinhood Ransomware Attack Targeting Baltimore

Date: May 28, 2025
Category: Ransomware / Data Breach

An Iranian national has acknowledged his involvement in a major ransomware and extortion operation linked to the Robbinhood ransomware in the U.S. Sina Gholinejad (also known as Sina Ghaaf), 37, along with his accomplices, infiltrated the computer networks of multiple U.S. organizations, encrypting files and demanding Bitcoin ransoms. Arrested in North Carolina in early January, Gholinejad pleaded guilty to charges of computer fraud and abuse, as well as conspiracy to commit wire fraud. He faces up to 30 years in prison, with his sentencing set for August 2025. The U.S. Department of Justice reported that these cyberattacks led to significant disruptions and financial losses exceeding $19 million for cities like Greenville, North Carolina, and Baltimore, Maryland.

ConnectWise Cyberattack: Nation-State Actor Suspected in Targeted Breach
May 30, 2025 | Vulnerability / Data Breach

ConnectWise, known for its remote access software ScreenConnect, has reported being targeted in a cyberattack believed to be orchestrated by a nation-state actor. On May 28, the company issued a brief advisory detailing that it had identified suspicious activity linked to the threat, which has affected a limited number of ScreenConnect customers. To investigate the incident further, ConnectWise has enlisted Google Mandiant for a forensic examination and has informed all impacted customers. While the company has not disclosed the specific number of affected customers, the timing of the breach, or the identity of the responsible party, it is important to note that just weeks prior, in late April 2025, ConnectWise addressed a high-severity vulnerability (CVE-2025-3935) with a CVSS score of 8.1 in ScreenConnect versions 25.2.3 and earlier.