22.09.2022
With AI pulling the trigger, a new generation of weapons of mass destruction looms. Thousands of researchers are now publicly declaring that they will not conduct research on autonomous weapons.
More than 2500 AI researchers and 176 organizations have signed a voluntary pledge not to conduct research on autonomous lethal weapons systems. Supporters of the initiative include prominent industry figures such as Demis Hassabis of Google DeepMind and Elon Musk. The core demand of the pledge is that firing a weapon should never be left to a machine.
By signing, participants pledge to "neither support nor assist in the development, production, trade, or use of autonomous lethal weapons."
So far, 176 AI-related organizations have declared their support, including major players on the scene. Many big names are missing, however; neither Google nor Google's holding company Alphabet signed the voluntary commitment; nor is Boston Dynamics - manufacturer of famous robot models - on the list so far. The relevant arms manufacturers are also missing.
The aim of the campaign is therefore to have an impact far beyond the circle of signatories by drawing the attention of the general public to the dangers of autonomous weapons. In this way, autonomous weapons are to be outlawed and companies researching their development are to feel public pressure.
In the long term, the initiators would like the international community to enact a ban on autonomous lethal weapons. Within the framework of the United Nations, 26 countries have already declared their support for such a project. Austria and the Vatican are the only European states on this list.
The commitment was initiated by the Future of Life Institute, a Boston-based organization that promotes projects that use technology for the benefit of all people and in the interest of future generations.
With their voluntary commitment, the initiators are not directed across the board against the use of artificial intelligence in a weapons system, but specifically against cases in which the computer alone decides whether or not the weapon is fired. They cite moral concerns against such systems - without a human decision-maker, the question of responsibility would not be clarified - but above all pragmatic considerations: Because, unlike humans, fully autonomous weapons systems can be multiplied at will, they would have a destabilizing effect on states and lead to a new arms race.
In November 2017, the institute had outlined how AI weapons could have an impact in an impressive video. In it, they describe the fictitious scenario of a terrorist group that uses mass-produced mini-drones to search for its opponents and execute them fully automatically.
