Google promises to no longer provide AI technology for weapons
Google came into the dark last week when it appeared that there had been an internal lie about the drone-deal that the company had concluded with the American Pentagon about the integration of image recognition via artificial intelligence into drones . Not only did the deal prove to be significantly more lucrative as had been said, nor was there any complete clarity about the form and length of the collaboration.
This is not
The company tries by means of Pichai to indicate where they want to go with artificial intelligence but fortunately they also indicate very clearly what they no longer want to cooperate from now on :
- Technology that is intended or has a risk of causing harm – they will be able to cooperate here in a very exceptional way, but only if it is proven that the result justifies the risks
- Weapons or other technology whose primary purpose is to is to cause human injury
- Technology that collects and / or supervises information beyond the internationally accepted norm to
- Technology whose goal is contrary to the accepted principles of international law and human rights in general
Good story
That is a clear list and it should be enough to temper the biggest riots. In addition, Google also immediately says that it does not mean that they will no longer cooperate with governments and military organizations. However, this will be in areas such as cyber security, training, recruitment and ‘search and rescue’.
Pichai closes the blog post with a promise to continue hearing other voices in the AI debate and to share the progress that the company is making in this field so that a multidisciplinary approach to the AI issue can be achieved thought about it. Good story, but the (leaked) collaborations will soon show whether Google will really stay that way.