Nội dung
- Military AI is advancing faster than verification can keep up
- Speed and scale could weaken the duty to protect civilians
- When AI does not just assist, but distorts human judgment
- Human Rights Watch’s recommendation: pause to control before broad deployment
- What this means for the Vietnamese market
- References
The use of artificial intelligence (AI) in life-and-death military decisions is fueling a debate that extends far beyond defense. For Vietnamese marketers, this is not just a technology or geopolitical story, but also a reminder of the responsibility that comes with applying AI: speed cannot replace judgment, and automation does not mean reliability. A new Human Rights Watch report shows that without proper controls, AI can lower the standard of human protection instead of raising it.
Military AI is advancing faster than verification can keep up
According to Human Rights Watch, military forces around the world are integrating AI into critical decisions such as target identification, the selection of force levels, and the assessment of risks to civilians. The problem is that deployment is happening far faster than systems can be tested, evaluated, verified, and validated.
The organization argues that the battlefield is being turned into a “real-world testing ground” for new technology. At the same time, AI is a class of systems that is very difficult to assess in traditional ways because of its “black box” nature, its dependence on simulations, and the fact that only part of its behavior can be tested. When a system is put into use without a sufficient understanding of how it works, the risk of violating international humanitarian law and international human rights law rises significantly.
Speed and scale could weaken the duty to protect civilians
The second notable point in Human Rights Watch’s warning is that AI is being used to plan and carry out attacks at a speed and scale that can make careful steps difficult, or even impossible. When the time available to verify targets and estimate civilian harm is compressed too much, citing the “speed of the machine” cannot be a reason to bypass mandatory precautionary measures.
In other words, if operational processes depend increasingly on algorithmic outputs, civilian protection standards may be pulled downward toward accepting probabilistic responses instead of legal assessment. This is an important risk not only for the military but for every industry accelerating automation: when processes are shortened, error and ethical consequences can also increase.
When AI does not just assist, but distorts human judgment
Human Rights Watch also warns about AI systems used for decision support. In theory, AI can improve user efficiency. But in military environments, this tool can weaken rather than support human judgment, which is the foundation for legal compliance.
The three issues highlighted are automation bias, which makes people trust machines more than they should; the difficulty of explaining machine-learning models; and the replacement of qualitative assessment with statistical inference. In addition, AI can generate false or misleading information while reproducing and amplifying existing patterns of discrimination. When these distortions are hard to detect, the final decision can easily be “validated” by the system’s output instead of being independently reviewed.
The core message of the report is: automation must not drag civilian protection standards down to a lower level simply because it is fast and convenient. With AI, the key question is not what the system can do, but how much real control humans still retain.
Human Rights Watch’s recommendation: pause to control before broad deployment
In its recommendations, Human Rights Watch calls on countries to immediately impose a moratorium on the use of any AI system, including generative AI, as the basis for targeting decisions. The organization also demands immediate transparency about AI’s role in the use of force and its impact on civilians and their rights.
The moratorium should only be lifted once minimum conditions are met. Those conditions include: not using data collected through human rights violations or unreliable data; ensuring that every system used in military operations has “meaningful human control”; preventing the de facto automation of life-and-death decisions; and requiring systems to be predictable, reliable, explainable, traceable, and transparent about how they are used and what impact they have.
The report also emphasizes that the use of AI must not lower the standard of civilian protection. This is a principle that matters far beyond warfare: technology only truly creates value when it improves decision quality and increases accountability, rather than shifting risk onto those affected.
What this means for the Vietnamese market
For businesses and marketing teams in Vietnam, this report is a direct reminder of responsible AI governance. As AI is increasingly used to analyze customer data, suggest content, automate customer care, or support decision-making, businesses also need to ask the same questions: is the data reliable, is the model explainable, who is responsible when the system makes a mistake, and do humans still retain the final review role?
In a market racing to adopt AI to speed up content production and optimize performance, sustainable value does not lie in “doing things faster at any cost,” but in building trust. Vietnamese businesses can see this as an opportunity to strengthen review processes, make AI usage more transparent, and put data safety standards ahead of short-term efficiency.
Source: Human Rights Watch, “Addressing Artificial Intelligence in the Military Domain” (14/6/2026).
See more marketing news and guides at https://marketing365.vn.
Follow more updates from Marketing365 to stay up to date with the latest marketing trends.
Read more articles in the same category Digital Trends.
This article focuses on AI in the military with a perspective for the Vietnamese market.
