USA wants to make one killer robot to test how dangerous it is

The group, led by former Google CEO Eric Schmidt, completed a two-day

public discussion of the use of AI to ensure national security and technological military progress.

Its vice-chairman, Robert Work, formerDeputy Secretary of Defense, said that autonomous weapons are expected to be less error-prone on the battlefield than humans. According to a Google representative, this will lead to a reduction in losses.

Over the past eight years, a coalition of non-governmentalorganizations insisted on concluding a treaty banning the creation and use of killer robots. The main argument of opponents of using AI on the battlefield is that human control is necessary to assess the proportionality of attacks and to blame war crimes. Thirty countries are calling for a ban on the use of killer robots according to the coalition website, and a United Nations body has met on the issue since at least 2014.

While autonomous weapon capabilities existFor decades now, concern has intensified with the development of AI to support them. In addition, numerous studies have uncovered biases in artificial intelligence systems and demonstrated examples of software abuse.

National Security CommissionThe US Artificial Intelligence Committee acknowledged the risks of autonomous weapons at meetings this week. However, the authorities still prefer to work to combat the spread of robotic systems based on AI, rather than introducing a corresponding treaty or a complete ban. The commission said that this “would be contrary to US interests.”

As a result, the commission members cameto the conclusion that Congress should at least think about giving killer robots or artificial intelligence systems a chance, Reuters reports. Otherwise, militaries may never know whether they are truly effective during military operations.

Mary Wareham, eight-year campaign coordinator"Stop Killer Robots," said the commission's emphasis "on the need to compete with similar investments made by China and Russia … serves only to encourage the arms race."

Beyond AI-powered weapons, the long reportthe commission was recommended to use AI by intelligence agencies to simplify the collection and analysis of data; allocate $ 32 billion annually in federal funding for AI research; create new bodies, including a digital corps modeled after the Army Medical Corps and a Technology Competitiveness Council chaired by the Vice President of the United States.

The Commission is due to submit its final report to Congress in March, but recommendations are optional.

Read more

Found a new kind of black hole that does not fit into the theory of relativity

Abortion and science: what will happen to the children who will give birth

Scientists have developed a replacement for the theory of relativity. What is the essence of the "theory of everything"?