Two documents published in September, one by former heads of state, one by a military affairs reporter embedded with Army chaplains, arrive at the same conclusion, but from opposite directions. The humans inside AI-enabled systems are not being adequately protected, and the institutions leading AI governance responsibilities know it.
The Elders published ten principles revealing that the governments of the world have failed to assume their AI governance responsibilities to protect their people from the existential and other threats posed by AI, as private companies and military systems expand AI’s influence over human rights, public safety, accountability, and decisions involving force.
The sentence doesn’t come from an academic paper or an activist manifesto but is the opening of a formal The Elders principles document released to the world on September 21st.
The Elders is the international organization of former heads of state and government founded by Nelson Mandela, whose current membership includes figures who have held the highest offices in their respective countries, who are, collectively, among the most experienced practitioners of sovereign governance alive.
AI’s Disruptive Societal and Economic Transformations
The Elders’ ten principles ask leaders to put humanity before profit, keep lethal decisions under human control, explain AI honestly, and cooperate internationally.
The Ten Principles the Elders have issues, Humanity, Safety, Cooperation, Trust, Accountability, Independence, Restraint, Solidarity, Integrity, and Long-term Thinking, can be perceived less like aspirational guidelines and more of an indictment of AI leaders’ current practices.
Together, they establish AI ethical principles for decisions affecting public services, infrastructure, employment, policing, borders, welfare, and war.
The Elders argue that AI is designed by companies, sold through markets, and subject to law. This makes AI governance responsibilities a government duty rather than one left to technology developers.
The international organization’s diagnosis is written in the language of obligation, not suggestion. One of the most disruptive societal and economic transformations of the 21st century is being sculpted by a small number – dominantly private – AI companies, such as OpenAI, Anthropic, Google, Meta, Microsoft, DeepSeek, Alibaba, among others, concentrated in two countries with systemic failure of oversight mechanisms.
The group says the US and China should begin negotiations on international safety standards because AI capabilities cross borders. Effective AI governance and compliance would enforce existing laws, create missing rules, and prevent excessive power from remaining concentrated among a few companies.
“AI must be governed to serve human beings, not to threaten their human rights or their very existence,” according to the The Elders paper.
That warning turns AI governance strategic visibility into more than knowing which systems an organization uses. Leaders must understand who built them, what data supports them, where people depend on them, and how benefits or losses are distributed.
For businesses, enterprise AI governance means assigning ownership before deploying tools in sensitive work.
AI governance monitoring should track errors, changing risks, security weaknesses, and unintended effects after deployment, not only during initial approval.
The Elders’ AI governance business reality means leaders cannot supervise systems they cannot examine, question, or stop.
AI Warfare Tests Human Control
On September 18, a small conference, Task & Purpose, convened within US Army Europe and Africa reported brought military chaplains to examine why the institution has never formally has to answer on what does a service member do with the moral weight of a killing they participated in, but did not meaningfully control?
Experts warned that faster AI-assisted targeting may leave troops responsible for lethal actions while giving them little meaningful control.
The 25th Infantry Division has already tested AI n a digital sensor-to-shooter artillery process that uses algorithms to optimize targeting data before soldier calculate firing solutions. So, AI adoption in military operations is no longer hypothetical, but has reached a point of no return where it’s now integrated into operational units.
“You could imagine moral injury arising from too little agency or from nominal responsibility without genuine control,” explained Jesse Kirkpatrick, a research professor at George Mason University who studies military ethics, autonomous weapons, and moral injury.
The psychological architecture of combat is changing much faster than the pastoral care infrastructure designed to support it.
Moral injury is psychological, moral, or spiritual harm caused by taking part in, witnessing, or failing to prevent actions that violate deeply held beliefs. It differs from post-traumatic stress disorder, although both can occur.
The battlefield example makes AI governance human validation critical. A human should not merely approve of a machine recommendation under pressure, that person needs information, authority, and time to challenge it.
The Elders’ restraint principle says decisions involving legitimate force must remain a human responsibility. It rejects weapons that select their own targets, placing AI governance responsibilities directly inside military command structures.
Pentagon policy requires appropriate human judgment over autonomous and semi-autonomous weapons. Yet policy cannot guarantee meaningful control when operators depend on AI-generated targeting information they may not fully understand.
This is where AI governance monitoring becomes necessary. Reviews should assess accuracy and whether a system’s speed, interface, or process increased pressure, reduced agency, or contributed to moral harm.
The Army’s response cannot rest only with chaplains. Experts said chaplains should work with mental health professionals, while operators need AI literacy to understand systems directly influencing their decisions.
For institutions and companies, AI governance continuous improvement requires evidence from users, affected communities, independent experts, and real incidents.
A matching approach to enterprise AI governance would update safeguards whenever new harms or dependencies appear.
Strong AI governance strategic visibility connects boardrooms, government agencies, and battlefields. Decision-makers must see both technical performance and human experience surrounding automated recommendations.
AI governance responsibilities become credible only when humans can pause systems, question outputs, document decisions, and support people carrying their consequences.
Long-term protection requires international rules, enforceable laws, transparent institutions, and AI governance continuous improvement. Without these protections, public promises may fail.
AI governance responsibilities must govern systems influencing decisions that people cannot inspect, contest, or reverse.
Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Ethical Tech section to stay informed and up-to-date with our daily articles.
