By Rev. Dr. Philippe SHOCK Matthews
The Thread This Month Has Been Following
We started this series with FORTRESS America, the Pentagon’s long-term initiative to make domestic bases resilient against extended infrastructure disruption. Then we turned to the question the Pentagon’s own planning raises for the rest of us: who is building comparable resilience for civilians? This week, I want to introduce the technology that sits quietly underneath both halves of that story. Artificial intelligence is becoming the system that predicts, defends, and increasingly manages the infrastructure that both the military and the civilian world depend on. That makes it one of the most important threads in this entire conversation, and one of the least understood.
AI Is Already the Operating Layer of Military Readiness
A military installation is not simply soldiers and buildings. It is an interconnected system of electrical power, water, fuel, transportation, communications, maintenance, and logistics, and that system has gotten too complex for human monitoring alone to catch every warning sign in time. The Pentagon’s technology office named Applied Artificial Intelligence and Contested Logistics among its six critical technology priorities, announced in November 2025 for fiscal 2026, reflecting exactly this shift (Office of the Under Secretary of Defense for Research and Engineering, 2025). The Department of Energy’s AI-FORTS initiative organizes its work around three pillars that map directly onto the FORTRESS logic: securing infrastructure from AI-enabled threats, securing infrastructure with AI, and securing the AI systems themselves (U.S. Department of Energy, 2026).
In practice, this means AI-enabled systems are increasingly used to forecast equipment failure before it happens, flag unusual network activity that might indicate a cyber intrusion, and help commanders estimate fuel and resource needs during a disruption. This is a real and useful capability. It is also, by its nature, a new dependency, and new dependencies are exactly where new vulnerabilities live.
A Technology That Defends and a Technology That Can Be Turned
Research on AI applied to critical infrastructure has found substantial gains in threat detection and response time. One study of critical energy infrastructure found threat-detection rates reaching 98% and incident response times cut by more than 70% in its evaluation (Govea et al., 2024). That is the case for AI’s defensive value, and it is a strong one.
But the same capability that detects anomalies can, if compromised, become the anomaly. An October 2026 report from the Government Accountability Office examined the Pentagon’s own oversight of its AI programs and found that the Responsible AI Working Council has no charter, and that only three of the 10 authorized positions in the department’s Responsible AI Division were filled as of February 2026 (U.S. Government Accountability Office, 2026). A separate GAO assessment, published in December 2024, found that most of the sector AI risk assessments sent to the Department of Homeland Security did not fully identify potential AI risks or how likely they were. GAO also noted that AI could improve the systems that run critical infrastructure like water and energy, and could make them more vulnerable to cyberattacks (U.S. Government Accountability Office, 2024).
Put plainly, and this is my reading of the findings: the government’s own watchdog is telling the government that it is adopting a powerful tool faster than it is building the guardrails for it. That gap between capability and accountability is not hypothetical. It is a documented, current one.
The AI Resilience Divide
Here is where I want to introduce a concept I believe deserves serious attention, and where my own research background in Africana phenomenology and the digital divide becomes directly relevant to a national-security story. Call it the AI Resilience Divide: unequal access to the AI-enabled systems that provide early warning, predictive maintenance, coordinated emergency response, and faster recovery during an infrastructure crisis.
The pattern behind it is not hypothetical. Research on tribal broadband access found that Native American tribes face structural barriers to the infrastructure investment required for digital resilience, and that responsibility for overcoming those barriers has often been shifted from federal institutions onto the communities themselves, obscuring how much of the gap is structural rather than a failure of local effort (Korostelina & Barrett, 2023). The same structural logic applies to AI-enabled resilience. A well-resourced community with redundant connectivity, predictive outage monitoring, and digitally coordinated emergency response experiences an infrastructure failure very differently from a community depending on a single provider, limited broadband, and alert systems that require a functioning cell signal to reach anyone at all.
Both communities can face the same triggering event — a grid failure, a cyberattack, a regional disaster — and experience meaningfully different consequences, not because one community is more careful or more deserving, but because one has been invested in and the other has not. That is a structural outcome, and it deserves structural scrutiny rather than being quietly absorbed as bad luck.
What Would Actually Happen if the AI Failed
It is worth being concrete about AI’s real limits here, because some of the public conversation treats AI as though it operates independently of the physical world. It does not. AI cannot generate electricity, purify water, manufacture a replacement part, or deliver food by producing a prediction. It depends entirely on functioning hardware, network connectivity, electricity, and human operators willing and able to act on what it reports. A system that loses power or connectivity loses its usefulness immediately, no matter how sophisticated its underlying model is.
This is precisely why real resilience — military or civilian — cannot be AI alone. It requires backup power, offline procedures that still work when the network doesn’t, trained human judgment that doesn’t depend on a dashboard, and clear lines of human authority that remain in place when the automated system goes quiet. An AI-enabled early warning system is genuinely valuable. It is not a substitute for the boring, physical redundancy — generators, water storage, trained people who know the manual procedure — that this series has already spent time on.
The Deeper Question
Through an Africana phenomenological lens, the question underneath all of this is not simply technical. It is a question about the distribution of protection. As national-security institutions invest heavily in AI-enabled infrastructure autonomy, we should be asking directly: will disparities in civilian access to energy, connectivity, and emergency technology create a new form of structural inequality: a differential capacity to survive and recover from the technologically mediated crises this series has been examining? That is a testable hypothesis, not a foregone conclusion. It deserves evidence-based, geographically specific research. That research should avoid assuming deliberate exclusion where unequal outcomes may instead reflect funding decisions, inherited infrastructure gaps, or geography. The consequences can be serious either way.
This is the reveal movement of the SHOCK Method™ applied at a structural, institutional scale rather than a personal one: clearly naming who is protected, who is not, and why, without assuming malice where the evidence points to neglect, and without excusing neglect simply because it isn’t malice.
Where This Leaves Us
The honest summary is this: AI genuinely strengthens the systems we depend on, and it genuinely introduces new dependencies and new gaps in oversight at the exact moment those systems are being asked to do more. Both things are true. Next week, in the final piece of this series, we turn to the deeper and more unsettling question this raises: what happens as AI systems themselves become more capable, more autonomous, and more trusted with consequential authority — and what does the actual scientific research say about whether that trust is, or could ever be, fully warranted?
Ready to go deeper?
Understanding structural inequality in who gets protected is research. Understanding how that inequality lives in your own body and nervous system is the work of healing it.
Book a Discovery Call: https://shockmetaphysics.gumroad.com/l/discoverycall
Learn the SHOCK Method™: https://shockmethod.com
Join the Patreon community: https://www.patreon.com/revshock
Nothing is wrong with Black people. Something happened to Black people. It’s time to break the Black trauma brain.
References
Govea, J., Gaibor-Naranjo, W., & Villegas-Ch, W. (2024). Transforming cybersecurity into critical energy infrastructure: A study on the effectiveness of artificial intelligence. Systems, 12(5), 165. https://doi.org/10.3390/systems12050165
Korostelina, K. V., & Barrett, J. R. (2023). Bridging the digital divide for Native American tribes: Roadblocks to broadband and community resilience. Policy & Internet. https://doi.org/10.1002/poi3.339
Office of the Under Secretary of Defense for Research and Engineering. (2025, November 17). Six critical technology areas [announcement, as reported by DefenseScoop]. https://defensescoop.com/2025/11/17/dod-six-critical-technology-areas-emil-michael-dow/
U.S. Department of Energy. (2026). AI-FORTS: Artificial Intelligence For Operationally Resilient Technologies and Systems. Office of Cybersecurity, Energy Security, and Emergency Response. https://www.energy.gov/ceser/artificial-intelligence-operationally-resilient-technologies-and-systems
U.S. Government Accountability Office. (2024, December 18). Artificial intelligence: DHS needs to improve risk assessment guidance for critical infrastructure sectors (GAO-25-107435). https://www.gao.gov/products/GAO-25-107435
U.S. Government Accountability Office. (2026, October 14). Artificial intelligence: Actions and opportunities to implement responsible use in DOD (GAO-27-107837). https://www.gao.gov/products/gao-27-107837
