The United States Attorney for the District of Maryland, Erek L. Barron, and the University of Maryland hosted the 2024 Cybersecurity Conference on November 7, 2024 in College Park, Maryland. Barron and the University of Maryland President, Darryll J. Pines, gave opening remarks. Barron proudly stated that the U.S. Attorney’s Office for the District of Maryland is working to “disrupt and dismantle the online infrastructure that facilitates cyber attacks.” Moreover, his office is also developing “investigations, prosecutions, and policies that complement and strengthen these disruptive efforts” on both the domestic and international scene. President Pines noted that “cybersecurity is one of the greatest challenges of our times.”
Several panels discussed the ongoing threats of emerging technologies, cybersecurity and generative artificial intelligence (AI). One of the panels specifically addressed AI threat assessments and responses, AI usage by adversaries, protection measures, and future threats. The panelists noted that AI makes it easier and cheaper to scale cyber attacks. And while deepfake video technology is improving, people can still detect deepfakes the substantial majority of the time. They have seen an increase in AI use for social engineering, making emails and voice calls more convincing. Thomas Sullivan, Deputy Chief, National Security and Cybercrime Section at the U.S. Attorney's Office for the District of Maryland, emphasized that victims of cyber crimes should report the attack as soon as possible to law enforcement. Sullivan also noted that law enforcement will not be able to prosecute away cyber threats. Mitigation is essential. Collaboration with service providers to prevent AI generated attacks is very important. Legislation is another angle for mitigation. The California AI Transparency Act which requires AI-generated content to be labeled is a great start, but additional regulation is necessary to protect the public. Incorporating AI-related risks into organizational policies and using multi-factor authentication is certainly the most immediate protection organizations can employ. Organizations should also create policies to handle AI and treat AI tools like cloud-based applications, ensuring proper controls are in place. Overall, the panelists underscored the importance of collaboration, regulation, risk assessment, and education in addressing the evolving threats posed by AI in cybersecurity.
The second panel discussed the risks and challenges of AI in financial transactions and the need for regulatory measures to mitigate these risks. The risks posed by AI include both facilitating and obscuring financial crimes, posing significant risks due to the interconnectedness of financial markets. One of the panelists, Ranganath Manthripragada, the Deputy Chief in the National Security and Cyber Crime Section of the U.S. Attorney’s Office for the District of Maryland, specifically emphasized the importance of tabletop exercises for risk management and having protective measures in place in all organizations. On a more personal level, Manthripragada warned about AI being used in scams like voice mimicry to exploit vulnerable individuals. In addition, Manthripragada indicated that AI poses a very high threat to national security, affecting critical infrastructure like the power grid and nuclear facilities. Near term challenges Manthripragada discussed were not waiting for incidents to happen before responding and encouraging a proactive, risk-based approach to regulation, like the European Union AI Act, which defines activity that is high risk and therefore more regulated. The panelists agreed that compliance efforts should align with protection measures, rather than being seen as mutually exclusive.
Overall, the panelists highlighted the dual role of AI as both a tool and a target, emphasizing the need for proactive measures, robust compliance, and international collaboration to mitigate risks. Other key points discussed were the ease in which AI makes criminal activity easier and harder to detect while also lowering the bar for entry into online deception. Consequently, there is a need to build security within AI products and perform risk assessments. Organizations should be developing policies on AI, and negotiating service agreements to ensure the third party provider is handling the organization’s data responsibly. Organizations may want to think about developing an internal large language model (LLM) so that all company proprietary information is used responsibly, sealed, and protected from disclosure. Auditing internal and external LLMs to ensure accuracy and accountability should be incorporated into any organization's policy regarding use of AI. Finally, don’t lose sight of the fact that humans need to be kept in the decisional loop.
_______________
The content of this article was taken from the
United States Attorney’s Office for the District of Maryland and the University of Maryland’s 2024 Cybersecurity Conference, November 7, 2024. Quotes were taken from Thomas Sullivan, Deputy Chief, National Security and Cybercrime Section at the U.S. Attorney's Office for the District of Maryland, and Ranganath Manthripragada, Deputy Chief in the National Security and Cyber Crime Section of the District of Maryland. Manthripragada is a career prosecutor, with over 34 years spent prosecuting cases in the state and federal criminal justice systems. He spent 20 years in the Counterterrorism Section of the United States Department of Justice, investigating and prosecuting terrorism cases both in the United States and abroad. While at the Department of Justice, Manthripragada was also involved in several international initiatives and was the National Security Division’s representative for the development of a multilateral solution to address the prosecution of Somali pirates. Manthripragada was also seconded to the Counterterrorism Bureau of the Department of State as a Senior Rule of Law Advisor, where he developed counterterrorism related training for prosecutors and judges worldwide in support of the Global Counterterrorism Forum. In addition to his international work, Ranganath was also the Critical Incident Coordinator in the National Security Division.