Since 1947, nuclear scientists have been setting a “doomsday clock” to alert the public to the current risk of nuclear war. The current state is 100 seconds to twelve, closer to twelve than ever before.
The risk of an accidental nuclear war exists above all, if an early warning system reports a missile attack in a crisis situation and if further events occur in a chronological context that can be put in relation to the warning message. The aspects described below will significantly increase the risk of an accidental nuclear war in the future.
Climate change will probably lead to different regions becoming uninhabitable and thus causing more climate refugees. The available habitat is becoming smaller, important resources such as water are becoming scarcer. As a result, political crises and possibly even armed conflicts will become more frequent in the future. As a result, missile attack reports will become much more dangerous. The increased risk of a nuclear war due to climate change has also been taken into account in the nuclear war clock since 2007.
Cyber attacks can create dangerous and unpredictable interactions with early warning systems and nuclear forces, significantly increasing the risk of accidental nuclear war. So far, there is little experience of possible cyber war processes. Cyber attacks are regarded as covert operations, war and peace are not clearly separated and may not be distinguishable. Military and civilian means can be combined and military and civilian targets can be attacked. A cyber war is probably difficult to control. It will also be difficult to limit a cyber war to individual states. Hacker groups that cannot be controlled by their states could interfere. The source of an attack can often not be determined. Although an attack can provide clues to its originator, it can also be deliberately misguided. There is the threat of an intensification of mutual attacks, which lead to an escalation spiral and can no longer be controlled by any side.
The new US military strategy permits an attack with nuclear weapons even in the event of a serious cyber attack. Even if there is hope that such a reaction will not normally take place, the situation will change if there is a missile attack report (as a false alarm) in an early warning system in a temporal context. However, such a military strategy also increases the risk of an incorrect assessment of an alarm message by potential opponents. If a missile alarm is received in temporal context with a cyber attack on the USA, a connection between these two events and the military strategy is obvious.
New arms race
The end of the INF contract will lead to a new arms race. In addition to the aspects of higher accuracy and shorter advance warning times, there are also indications of plans for smaller nuclear weapons that could be used sooner. An assumed lower operational threshold also increases the danger that an alarm message is assumed to be valid, as the use of nuclear weapons is more likely. This also increases the risk of an accidental nuclear war.
The effects of the planned arming of outer space and the development of hypersonic weapons, which are apparently difficult to locate and which will extremely shorten early warning times, are completely incalculable. This will drastically increase the complexity of possible alerts, and it will hardly be possible to test early warning systems for these new weapons.
The number and variety of objects in airspace will continue to increase (e.g. drones, satellites, hypersonic missiles). The evaluation of sensor signals will thus become more difficult and more and more artificial intelligence (AI) methods will be required to automatically make decisions for certain subtasks. The further development of weapon systems with higher accuracy and shorter flight times will also increasingly require artificial intelligence techniques. There are already demands to develop autonomous AI systems in connection with early warning systems, as there is no time for human decisions.
However, testing such systems under real conditions is hardly possible. Compared to other AI applications (e.g. autonomous driving), there will also be significantly less “learning data” to generate the necessary recognition criteria. This can lead to unpredictable effects that may not be evaluated and controlled by humans. In the short time available, it will usually not be possible to check machine decisions. The human being can only believe what the machine delivers.
Many states with nuclear weapons
There are now several states with nuclear weapons. Not only the USA and Russia have early warning systems, but also states such as China are building them up. These early warning systems can also lead to mistakes and wrong decisions with fatal global consequences.
War => Nuclear War
If it comes to a military conflict in which states with nuclear weapons are involved in opposition, then most likely a “hybrid warfare” with increasing cyber attacks will follow. Because it is not very plausible that in such a case no cyber attacks will be used. But a cyber war can easily get out of control, especially since there is no experience with an extensive cyber war so far. Errors in early warning systems with reports about attacking nuclear missiles are very dangerous in such situations. If, due to a computer error during a war conflict, an early warning system reports nuclear missiles that are attacking and if this is related to other events, e.g. serious cyber attacks, a causal connection is plausible for the evaluation teams. This means that there is a danger that the report of an attack will be accepted as valid, which can lead to an accidental nuclear war. Such a danger exists today not only between USA and Russia, but can induced ba any nation having nuclear weapons, e.g. in the context of a conflict between India and Pakistan. Even a limited nuclear exchange, e.g. between India and Pakistan, can lead to a nuclear winter with serious consequences for the humanity as a whole.
Translated with www.DeepL.com/Translator