Artificial intelligence is improving U.S. special operations while making digital information harder to trust, a senior commander said at a security conference in Honolulu. Adm. Frank Bradley, leader of U.S. Special Operations Command, warned that increasingly convincing deception could weaken the coordination needed to deter or respond to an attack. Defense One reported his remarks.
Bradley said hostile actors can now use widely available AI tools to create deepfakes and other persuasive material. He said that capability is undermining military and civilian leaders’ confidence in digital information and raising doubt about what is and is not true. That loss of credibility, he said, could affect the ability of the United States to generate unity of effort to deter or defend against an adversarial attack.
At a Global SOF Foundation symposium on Indo-Pacific irregular warfare, Bradley described how capable AI-generated content can be. He said people outside an insider group might not be able to distinguish AI-created material from authentic content. The examples discussed included synthetic voices, cloned faces and complete false identities.
A U.S. military official working on counterintelligence and foreign influence analysis said AI has made deception scalable. The official said that, beyond individual campaigns, the technology could undermine confidence in intelligence gathering and make every source of information increasingly suspect.
Ongoing Russian and Chinese efforts to deceive Western audiences, including efforts intended to influence public sentiment and gain tactical advantages, are receiving a boost from AI tools, according to the report. The report also described the use of AI tools by Ibrahim Traoré, the leader of a junta controlling Burkina Faso, as an example of how smaller-scale authoritarian governments can use the technology to influence populations.
Bradley said more persuasive AI-enabled content, used quietly and consistently to shape perceptions about events and locations, has become a central feature of a new information battleground. He described exposing hostile behavior accurately and early as a critical advantage, adding that information advantage leads to decision advantage.
A former State Department intelligence official said the United States had surrendered tools used to watch and detect foreign influence while adversaries were expanding their ability to conduct influence operations. The official said the information environment was less transparent than it had been in the past, citing changes in leadership at social-media companies and reduced transparency around content moderation and disinformation.
The former official also referred to government offices and programs involved in monitoring foreign influence, including the FBI’s Foreign Influence Task Force, the Office of the Director of National Intelligence’s Foreign Malign Influence Center and various USAID activities. The official said it was unclear who was able to see the overall picture.
AI is also giving U.S. forces new ways to combine information and conduct operations. Bradley pointed to the April rescue of a downed F-15 pilot behind enemy lines, which he said was made possible by synchronized intelligence, cyber-enabled support and disciplined coordination across multiple elements.
He said operations of that kind require a shared intelligence picture that changes as quickly as real-world conditions and data that can effectively predict how those conditions will change. An intelligence official told Defense One that systems once reserved for specialized analysts now combine dozens or even thousands of data sources for operators, with updates every 15 minutes.
The officials said the challenge is not only finding new intelligence and data streams but also trusting them. Bradley said adversarial AI agents are finding new methods to access and alter data and intelligence.
He said the U.S. military and other parts of the federal government are working to ensure that datasets and systems are appropriately protected against agentic AI attacks that might be coming. The information environment, Bradley said, is now less transparent than it was in the past, and AI will make it worse.