The United States and Russia have relaxed rules for combat AI: machines may be given more freedom to decide whom to attack

 • 6476 переглядiв

The draft UN rules removed the mandatory human verification of a target before a strike. This could increase the risk of mistaken attacks and civilian casualties.

The United States and Russia succeeded in weakening a number of proposed restrictions on the use of artificial intelligence in lethal autonomous weapons. In particular, the requirement that a human verify an AI-selected military target before a strike was removed from the document, The Washington Post reports, writes UNN.

Details

The issue concerns UN negotiations on future international rules governing the use of autonomous weapons. According to the publication, they represented the biggest progress in recent years toward a potential global agreement in this area.

However, during the final stage of the negotiations, the U.S. and Russian delegations secured the removal of several safeguards at once. In particular, the requirements that systems operate "predictably" and "reliably" were removed from the text, as was a provision on taking into account the ethical aspects of using AI in combat.

What threat could this pose to people

The greatest concern was caused by the removal of a provision requiring a human to verify a military target identified by artificial intelligence before a strike was carried out.

Leaders of leading AI companies warned the UN about the risks of artificial intelligence to humanity24.09.26, 11:06

In simple terms, this concerns the risk of the emergence of combat systems in which an algorithm could independently find and select a target, while human control over the final decision would be weaker or not guaranteed at all by international rules.

This could mean that machines would be able to make life-and-death decisions without human control

– said Verity Coyle, a representative of Human Rights Watch.

According to her, the consequences could include a greater number of civilian casualties, difficulties in determining who is responsible for an erroneous strike, and a faster transition to automated warfare.

The particular danger lies in the speed of such systems. If states deploy autonomous weapons on a mass scale, algorithms could potentially detect targets and respond much faster than humans. In the event of an error, misidentification, or incorrect data, there may be significantly less time for human intervention.

British companies will use battlefield data from Ukraine to develop AI-powered drones25.09.26, 14:49

Popular
Taylor Swift shared rare footage from a trip to Travis Kelce’s hometown

 • 6354 переглядiв

Sylvester Stallone confessed to battling bulimia and using steroids at the peak of his career

 • 10488 переглядiв

In Germany, the routes of around 300 trains were disrupted due to the theft of railway cables

 • 3804 переглядiв

News by theme
Grisly details of Putinist Sergei Sosedov’s death have emerged

 • 1986 переглядiв

Occupiers attacked a car in the Zaporizhzhia region: a man and a woman were killed

 • 1828 переглядiв