The Department of Homeland Security has released a report exploring the technologies used to develop and propagate digital content forgeries and how foreign governments and nongovernmental organizations can use such technologies to weaponize such forgeries.
DHS said Friday the report discusses the potential risks and harms of artificial intelligence-synthesized fake media on homeland security missions.
The 99-page document, titled Impacts of Adversarial Use of Generative AI on Homeland Security, also describes risk-reduction strategies to mitigate risks from digital forgeries.
Key Technical Observations and Considerations
The report presents eight key technical observations and considerations: convenience and security; multipronged mitigation approaches; threats of tomorrow; AI software; AI hardware and the AI race; technical intelligence; artificial general intelligence milestones; and whole of government responses.
For multipronged mitigation approaches, the report offers some key observations relevant to the design of a risk reduction strategy.
These include the need for other methods to prevent the spread of deepfakes and the role of human training and automated contextual analysis tools in detecting deepfakes.