Articles of Interest: April 12-18, 2025
- The Prometheus Security Project Team
- Apr 18
- 5 min read
Tricia Crimmins, Tech Brew. April 16, 2025.
The DoD is partnering with private companies to build nuclear microreactors on military bases through its Advanced Nuclear Power for Installations (ANPI) program. Eight companies, including submarine nuclear supplier BWXT Advanced Technologies, have been selected as eligible partners. Despite this program being started during the Biden administration it continued to fall under the current administration's energy goals and executive orders. The DoD is framing this partnership as enhancing military energy resilience and reducing dependence on foreign energy supplies.
The Pentagon is effectively acting as both a customer and investor in advanced nuclear technologies that may have struggled to gain traction in the private sector. Military bases functioning as test beds for next-generation nuclear technologies could provide crucial operational data and pathways to public acceptance that would benefit civilian applications. For everyday Americans, this military-first deployment strategy means their first exposure to advanced nuclear energy might come through its integration into defense infrastructure rather than through civilian power grids or commercial applications that have concerned people in the past. Energy technology breakthroughs will become increasingly important as other emerging technologies like AI continue to advance and demand even higher amounts of energy to produce. The DoD taking this step toward investing in advanced energy research may prove to be an important step toward energy breakthroughs that benefit the whole of our country, not just the DoD.
Crimmins, Tricia. "DoD invites companies to build nuclear reactors to power military bases." Tech Brew, April 16, 2025. http://emergingtechbrew.com/stories/2025/04/16/defense-department-nuclear-reactors
Ty Roush, Forbes. April 16, 2025.
A bipartisan House committee report claims that Chinese AI startup DeepSeek poses a "profound threat" to U.S. national security. The report argues that DeepSeek harvests user data for the Chinese government, relies on stolen U.S. technology, and manipulates search results to align with Chinese propaganda. The committee also suggests DeepSeek's founder is tied closely with the Chinese government. While the report doesn't recommend banning DeepSeek outright, it proposes new export controls and prohibiting federal use of AI models produced in China. Several U.S. agencies and other countries have already restricted DeepSeek access on government devices due to security concerns.
As AI becomes increasingly viewed as a security tool rather than a shared scientific advancement, international cooperation becomes more difficult and thus norm building and communication to avoid misunderstandings become more difficult. For everyday Americans, this means the AI tools and services they use may increasingly be determined by geopolitical tensions rather than by technological merit or consumer choice.
By framing AI innovation primarily as a security competition, we risk creating a fragmented global AI landscape where differing regulatory approaches form camps of support further perpetuating an arms race mentality of “us vs them.” China’s approach will differ from the U.S. approach which will differ from the European approach and so on. Differing priorities, differing values, and differing lines each camp is willing to cross will almost inevitably cause tension and make collaboration on topics like AI safety and AI governance more difficult.
This security-first approach also threatens to overshadow important discussions about how AI should be governed to benefit humanity broadly. It also suggests that future AI innovation may become increasingly classified or restricted, limiting the transparency needed for proper public oversight and ethical development of these powerful technologies.
Roush, Ty. "China’s DeepSeek AI Poses ‘Profound Threat’ To National Security, House Report Claims." Forbes, April 16, 2025. https://www.forbes.com/sites/tylerroush/2025/04/16/chinas-deepseek-ai-poses-profound-threat-to-national-security-house-report-claims/
Matthias Monroy, Matthias Monroy (blog). April 14, 2025.
Thanks to aggressive political pressure from certain factions within the European Parliament, the EU has enhanced its border monitoring capabilities. The EU’s border protection agency, Frontex, plans to increase its UAV fleet with newly purchased drones. This will help Frontex expand its monitoring of the Mediterranean Sea, but also lead to further ethical questions. Will Frontex's expansion of its UAV force lead to migrants being helped at sea? Or will it lead to more aggressive deportation tactics?
In 2024, Frontex ramped up its UAV flying time by several thousand hours. This was in part because of political pressure throughout the EU and the continued flow of migrants into Europe’s borders. Now, Frontex wants to upgrade its fleet by buying new UAV models from an Israeli manufacturer. The Heron I UAV is now one of the main drone systems being used by Frontex. Thanks to recent technological developments, the Heron I has newly designed cameras and an increased range. Frontex’s use of these drones has enabled it to monitor the Mediterranean for longer periods of time.
The imagery gained from the Heron I helps Coast Guard units sail towards any boats headed to Europe. However, Frontex also sends information gained by drone surveillance to North African countries, whose security agencies are regarded as brutal and violent towards migrants. As a result, Frontex faces a conundrum: should it provide information gained from drone surveillance to countries with atrocious human rights records like Libya? This debate is mirrored in the US, which routinely requests that other Latin American countries assist in clamping down on migration flows northward. If more drone and surveillance technology is used to assist in cracking down on immigration, then in what ways does that violate the right to seek asylum, or does it lead to regimes gaining access to Western drone technology in order to crack down on people fleeing from human rights abuses?
Monroy, Matthias. “Frontex Significantly Expands Drone Operations: Number of Flight Hours Increases by 50 per Cent.” Matthias Monroy (blog), April 14, 2025. https://digit.site36.net/2025/04/14/frontex-significantly-expands-drone-operations-number-of-flight-hours-increases-by-50-per-cent/.
Emanuel Maiberg, Jason Koebler. 404 Media. April 17, 2025.
Through its product Overwatch, New-York based company Massive Blue is selling and providing services using AI and virtual agents to aid police departments in finding criminals. The product creates a realistic, yet fake, AI persona to engage with others online.
The use of this AI tool may seem helpful at first, but many things could go wrong very quickly without the proper policies or parameters set. There is already an issue in the United States with the revoking of student visas over protesting, and further discrimination. Pinal County Sheriff's Office public information officer Sam Salzwedel claims “Massive Blue is not working on any immigration cases” and “is focused on the areas of human trafficking, narcotics trafficking, and other investigations.” However, if this product and its programming is not carefully monitored and regulated, it may start to target immigration cases which could lead to improper profiling of those deemed criminals and more unlawful deportations. There is space for human error, and thus also exists space for AI to make technical errors as well.
Maiberg, Emanuel and Koebler, Jason. "This ‘College Protester’ Isn’t Real. It’s an AI-Powered Undercover Bot for Cops." 404 Media, April 17, 2025. https://www.404media.co/this-college-protester-isnt-real-its-an-ai-powered-undercover-bot-for-cops/
