Articles of Interest: April 5-11, 2025
- The Prometheus Security Project Team
- Apr 11
- 4 min read
Shaun Waterman, Air and Space Forces Magazine. April 9, 2025.
The Pentagon's first AI chief, retired Lt. Gen. John Shanahan, warns that employing autonomous military systems requires overhauling both tactics and testing processes. Unlike traditional software with scheduled updates, AI systems need continuous testing and updating in order to maintain its effectiveness and to counter adversaries' attempts to “corrupt” it. The military must develop new testing approaches that can be implemented at the unit level during conflicts while ensuring proper integration with other systems and addressing the unique challenges posed by autonomous decision making.
As autonomous systems become more integrated into the joint forces and essential to battlefield strategy, the traditional "test then deploy" model becomes dangerously outdated. The race against competitors like China creates pressure to accelerate the deployment of these systems, potentially at the expense of thorough evaluation.
When autonomous systems interact or engage in complex environments with each other, with humans, and against adversary systems, new and unforeseen risks will likely emerge that could have catastrophic consequences for military personnel, potentially costing lives.
This balancing act of speed vs safety is seen at every level of AI innovation. At the federal scale we see the current administration moving away from safety regulation in favor of innovation speed, while at a smaller scale issues like the testing and deployment of autonomous systems in a military context must make a similar choice. Will thorough safety and testing be overlooked in favor of rapid battlefield deployment? Or will we armed forces take a cautious approach to these new systems, potentially risking falling behind adversaries who chose speed rather than safety? Perhaps there is a middle ground where both safety and speed can be achieved. Ultimately, AI is going to continuously face the question of speed vs safety, how we decide to answer that question will likely define the future.
Waterman, Shaun. "Military AI Will Mean Overhauling Test as Well as Tactics: DOD’s First AI Chief." Air and Space Forces Magazine, April 9, 2025.. https://www.airandspaceforces.com/military-ai-overhauling-test/
Emeka Anuforo, African Development Bank Group. April 4, 2025.
On April 3rd, the African Development Bank, along with the government of Mozambique and South Korea’s Busan Technopark, announced the creation of a program to train drone operators in Mozambique to monitor flood damage in areas of the country affected by chronic flooding. Busan Technopark will help finance the project over a period of six months.
As the likelihood of extreme weather patterns increases, countries at risk of natural disasters have been exploring how to react. The use of UAVs to monitor disaster zones is a crucial way to enable governments to respond more quickly. In Mozambique’s case, flood damage affects much of the country’s critical infrastructure, hindering a disaster response plan. By using an array of UAVs, the government can identify the worst-hit areas more quickly and react accordingly to save lives and property. While the article does not identify which type of UAV will be used, it is assumed that the equipment will be cheaper than using aircraft or helicopters to monitor a disaster zone.
The use of UAVs to assist with disaster response is an important step in using emerging technology to assist in times of natural disasters. Small UAVs provide governments with the ability to identify a crisis area and assist in responding. People caught in a flood-affected zone can be saved more quickly with drones spotting them from above. If more countries experiencing similar weather patterns as Mozambique employ UAVs as a monitoring tool, then disaster response can be sped up and more lives will be saved.
Anuforo, Emeka. “African Development Bank and Mozambique Launch Drone-Based Initiative to Strengthen Country’s Disaster Preparedness.” African Development Bank Group, April 2025. https://www.afdb.org/en/news-and-events/press-releases/african-development-bank-and-mozambique-launch-drone-based-initiative-strengthen-countrys-disaster-preparedness-82457.
Luke Cox, Science. April 3, 2025.
Within the pharmaceutical industry, most of the work is done manually. Researchers spend countless hours working through complex processes to grow and maintain cells. It is difficult, time-consuming, expensive and comes with contamination risks.
Scientists at the University of Bristol have created a method of moving cells without needing to physically touch them. Instead, acoustics are used to make the cells “dance” and move around their containers. This technology will hopefully open the door for cheaper and more accessible automation. Instead of requiring enough equipment to fill a van, the same work can be done on a small work bench.
Automation typically increases productivity, and with cell research remaining a major focus of the pharmaceutical industry, there is hope that medications could be created and tested at a much faster rate and for a fraction of the cost.
Cox, Luke. "Dancing with the cells: How acoustically levitating a diamond enabled a redesign of biotech automation." Science, April 3, 2025. https://www.science.org/doi/10.1126/science.adw0393
Nathan Kahl, George Mason University. March 31, 2025.
After approval from the State Council of Higher Education for Virginia (SCHEV), effective fall 2025, George Mason University will be offering a master of science in artificial intelligence (AI). The courses will offer an accessible foundation to AI, and the program has the goal of positively influencing the future of AI.
Considering the quick and ever changing nature of AI progression, the question of ethics is brought up quite often. There is a need for accountability and responsibility in the tasks and training that is programmed into AI systems. The implementation of a master’s degree dedicated to AI by George Mason University is a step in the right direction to educating and preparing a future generation on how to ethically engage with AI. Other academic and professional avenues that could be influenced to expand are the legal and policymaking fields, which could expand to have more concrete laws and policies that are implemented to protect individuals and companies against abuse from AI.
Kahl, Nathan. “George Mason launches Virginia’s first public master’s degree in AI.” George Mason University, March 31, 2025.
