AI can already assist humans with tasks involving perception, analysis, prediction, logistics and decision support. Military organizations around the world are interested in AI because it can process enormous amounts of information and potentially help people make decisions more efficiently.
That capability can have legitimate defensive and humanitarian applications.
For example, AI can potentially help analyze large datasets, identify patterns, improve logistics, support planning or assist with disaster response.
But the same general technology can raise serious concerns when it becomes connected to systems capable of causing physical harm.
The United Nations has repeatedly emphasized the importance of maintaining human control over decisions involving the use of force. In May 2025, Secretary-General António Guterres called lethal autonomous weapons operating without human control “politically unacceptable” and “morally repugnant,” while urging countries to work toward legally binding international rules.
That position highlights a fundamental principle:
A machine may process information, but humans must remain accountable for life-and-death decisions.
This is significantly different from the fictional premise of The Terminator.
Skynet represents a machine independently deciding humanity's fate.
Modern concerns are more immediate: humans could gradually delegate more responsibility to automated systems because those systems appear faster, cheaper or more efficient.
The danger could therefore emerge not from a machine taking power overnight, but from people slowly giving away control.
Why Speed Could Become a Problem
One of Cameron's most striking concerns is the speed at which automated systems could operate.
Humans need time to perceive information, interpret it, communicate with others and make judgments. Computers can process information far more rapidly.
That difference can be useful when the task is routine.
But in a crisis, speed can also create danger.
Imagine a situation in which multiple automated systems respond to rapidly changing information. One system interprets an event as a threat. Another reacts to that assessment. A third system receives the new information and updates its own calculations.
If each stage happens faster than humans can meaningfully evaluate it, the decision-making process can become increasingly difficult to interrupt.
This is sometimes described as an automation problem: once humans begin relying heavily on automated decisions, they may have difficulty recognizing when the system is wrong.
The issue is not that computers necessarily “want” to make a particular decision.
The issue is that a system can be wrong while appearing confident.
That distinction is especially important for AI because advanced models can produce outputs that look convincing even when their reasoning or conclusions are flawed.
In civilian applications, an incorrect AI recommendation might cause inconvenience or financial loss.
In a high-stakes military context, the consequences could be dramatically more serious.
Governments Are Already Trying to Keep Humans in the Loop
Cameron's warning does not mean governments are simply allowing AI systems to operate without restrictions.
In fact, some governments have explicitly created policies intended to preserve human judgment.
The U.S. Department of Defense updated its directive on autonomous weapon systems in 2023. The policy states that autonomous and semi-autonomous weapon systems should be designed so that commanders and operators can exercise appropriate levels of human judgment over the use of force.
The same directive emphasizes responsible and lawful use, testing, reliability and safeguards intended to reduce the likelihood and consequences of unintended engagements.
The Department of Defense has also established broader ethical principles for military AI, including requirements related to responsibility, traceability, reliability and governance. Its framework stresses that people must remain responsible for the development, deployment and use of AI capabilities.
That is an important contrast with the world of The Terminator.
In Cameron's fictional future, the machine ultimately becomes the decision-maker.
In responsible AI frameworks, the objective is supposed to be the opposite: AI remains a tool within a human-controlled system.
The challenge is making that principle work in practice.