> ## Content Index > Fetch the complete content index at: https://www.thedigitalspeaker.com/llms.txt > Use this file to discover other available public pages before exploring further. # When Your Next Online Friend is a Pentagon-Crafted Illusion - URL: https://www.thedigitalspeaker.com/when-your-next-online-friend-pentagon-crafted-illusion/ - Published: 2024-10-21T06:04:14.000Z - Updated: 2026-08-04T05:37:32.000Z - Description: The Pentagon is hard at work developing AI-generated personas so convincing that neither humans nor AI will be able to tell them apart. In an era already rife with deepfakes and misinformation, we must ask: what could possibly go wrong? - Author: Dr Mark van Rijmenam, CSP - Tags: News, #seo-post-1 In a world where your next best friend might not even be human, the Pentagon is now working to create fake online personas so convincing, even [AI](https://www.thedigitalspeaker.com/ai-strategy-speaker/) won’t know they’re not real. Sounds like the plot of a dystopian novel, right? Welcome to the future of online deception. - The Pentagon plans to create AI personas indistinguishable from real people, using technologies like StyleGAN. - These personas will not only feature static images but also realistic selfie videos, making them nearly impossible to detect. - The U.S. military’s use of this technology conflicts with its previous warnings about the dangers of deepfakes. In a startling development, [the U.S. military’s Joint Special Operations Command (JSOC) is on a mission to build AI-generated online personas that are virtually indistinguishable from real people](https://www.documentcloud.org/documents/25224425-jsoc-stylegan-personas?ref=thedigitalspeaker.com). These fake personas—complete with government ID-quality photos, multiple facial expressions, and convincing background environments—are being developed as part of covert operations to gather intelligence on [social media](https://www.thedigitalspeaker.com/digital-ethics-speaker/) platforms. The technology behind this initiative includes cutting-edge systems like Nvidia’s StyleGAN, which is capable of generating hyper-realistic digital faces that look entirely human. ## AI Selfie Videos But these avatars aren’t just limited to still images. [JSOC aims to go further](https://theintercept.com/2024/10/17/pentagon-ai-deepfake-internet-users/?ref=thedigitalspeaker.com) by creating "selfie" videos where fabricated individuals interact in virtual environments, communicating with AI-synced audio to complete the illusion. These personas would be used to infiltrate public forums and online networks, gathering intelligence without raising alarms. Imagine chatting with someone on a platform, sharing ideas, only to realize months later that your “friend” never existed. While the technology sounds like something out of a sci-fi thriller, its real-world implications are profound. The [U.S. military has openly criticized](https://cyberscoop.com/foreign-ai-use-isnt-fooling-masses/?ref=thedigitalspeaker.com) the use of [deepfake technology](https://www.thedigitalspeaker.com/how-fight-deepfake-scams/) by adversarial states such as Russia and China, which have used similar tools for [disinformation](https://media.defense.gov/2023/Sep/12/2003298925/-1/-1/0/CSI-DEEPFAKE-THREATS.PDF?ref=thedigitalspeaker.com) and influence campaigns. Yet, the Pentagon’s move to develop these [AI-driven deepfakes](https://www.thedigitalspeaker.com/rise-deepfakes-digital-reality-becomes-fake/) introduces a troubling double standard. What happens when [the lines between truth and fabrication blur](https://www.thedigitalspeaker.com/how-to-trust-each-other-in-a-post-truth-world/) to the point where even advanced AI can’t detect the difference? The technology itself is a significant leap forward in AI manipulation. [StyleGAN](https://www.thedigitalspeaker.com/gans-limited-data-synthetic-content-generation-ai-impact-business/), originally released in 2019, became famous for powering the popular website "This Person Does Not Exist," where each refresh of the page generates a new, lifelike face of a person who never existed. The U.S. military, however, is aiming far beyond static images. Their ambition extends into dynamic environments, where entire virtual lives could be fabricated—complete with selfie videos and matching audio that deceive even the most sophisticated detection systems. Ethically, this raises enormous concerns. [A government that has previously condemned the use of AI-generated content](https://media.defense.gov/2023/Sep/12/2003298925/-1/-1/0/CSI-DEEPFAKE-THREATS.PDF?ref=thedigitalspeaker.com) to manipulate public opinion is now leaning into the very same technology for its own clandestine operations. [According](https://theintercept.com/2024/10/17/pentagon-ai-deepfake-internet-users/?ref=thedigitalspeaker.com) to Heidy Khlaaf, chief AI scientist at the AI Now Institute, “There are no legitimate use cases besides deception.” If the Pentagon succeeds in creating these undetectable fake identities, it could normalize the use of [deepfake personas for intelligence and warfare](https://www.thedigitalspeaker.com/ai-dark-side-how-criminals-use-ai/)—paving the way for authoritarian states to follow suit. This technology could undermine public trust, both domestically and globally. Will citizens start to question the authenticity of every online interaction (as they already should)? Moreover, will foreign adversaries retaliate by developing even more advanced tools of deception, escalating an arms race where AI-driven disinformation becomes the norm? As AI-driven deception becomes more prevalent, the question isn't just about what we can build—but whether we should. In a world where trust is already fragile, do we really want to live in a digital landscape where even the people we interact with online may not exist? Read the full article on [The Intercept](https://theintercept.com/2024/10/17/pentagon-ai-deepfake-internet-users/?ref=thedigitalspeaker.com). \---- ## Frequently asked questions ### What is JSOC doing with AI-generated personas? JSOC, the U.S. military's Joint Special Operations Command, is developing AI-generated online personas that are virtually indistinguishable from real people, complete with government ID-quality photos, multiple facial expressions and convincing background environments. These fake identities are intended for covert operations to gather intelligence on social media platforms without raising alarms among users. [Link to this question](#faq-what-is-jsoc-doing-with-ai-generated-personas) ### How does StyleGAN factor into these fake personas? StyleGAN, a system created by Nvidia and originally released in 2019, generates hyper-realistic digital faces that look entirely human. It became famous for powering the website This Person Does Not Exist, where each refresh produces a new lifelike face of someone who never existed. The Pentagon is using this same technology as the foundation for its fabricated online personas. [Link to this question](#faq-how-does-stylegan-factor-into-these-fake-personas) ### Why is the Pentagon's use of deepfakes considered hypocritical? The U.S. military has openly criticized adversarial states like Russia and China for using deepfake technology in disinformation and influence campaigns. Yet the Pentagon is now developing the same kind of AI-driven deepfakes for its own covert operations, creating a troubling double standard between condemning manipulative AI content abroad while embracing it for domestic intelligence purposes. [Link to this question](#faq-why-is-the-pentagon-s-use-of-deepfakes-considered) ### What are the risks of undetectable AI personas becoming normalized? If the Pentagon succeeds in creating undetectable fake identities, it could normalize deepfake personas for intelligence and warfare, encouraging authoritarian states to follow suit. This risks undermining public trust both domestically and globally, making people question the authenticity of online interactions, and could trigger an arms race in which foreign adversaries build even more advanced disinformation tools. [Link to this question](#faq-what-are-the-risks-of-undetectable-ai-personas-becoming)