Should robots be given legal status? This is the question that currently occupies the AI experts, and for a good reason, a paragraph well buried in a report of the European Parliament published in January 2017, suggests that it is necessary to assign “electronic personalities” to certain robots.
“The creation, over time, of a legal personality specific to robots, so that at least the most sophisticated autonomous robots can be considered as electronic persons responsible for repairing any damage caused to a third party; it would be conceivable to consider as an electronic person any robot that makes autonomous decisions or that interacts independently with third parties.”
Translation, this recommendation suggests that autonomous robots should be held accountable, especially in the event that these machines damage the safety of people or damage property.
The parties to this proposal, including builders, claim that this change is common sense. Legal status will not give robots the right to marry or to enjoy the same rights as humans. Far from it, it will equalize the AI with companies, which they already enjoy the status of “legal person,” and are treated as such in courts around the world.
But these arguments do not convince the opponents of this proposal. In a letter to the European Commission, 156 experts in artificial intelligence from 14 European countries, computer scientists, law professors and CEOs warned Brussels of the danger of implementing this proposal. According to them, assigning a legal status to robots would be “inappropriate” and pose “legal and ethical” problems.
On the side of the legislative authorities, Mady Delvaux, Vice-President of the Judicial Committee of the European Parliament and Luxembourg Socialist MEP, said that although she is not sure whether granting a legal personality is a good idea, it is increasingly convinced that the current legislation is limited and is unable to deal with the complex problems that concern the responsibility of autonomous machines. For this reason, she thinks that all options need to be explored.
A position categorically rejected by AI experts in their petition and for whom this proposal is almost heresy.
Noel Sharkey, Professor of Artificial Intelligence and Robotics at the University of Sheffield, also agrees that for builders, this proposal is a way to get rid of responsibility for the actions of their machines and the damage they do could cause.
Robots with intelligence similar to humans and capable of making decisions are still far from existing. Today, robots do better than humans in certain specific tasks, even limited applications, such as image recognition, or play the game of Go.
But these applications, although advanced, remain limited to a single domain. Playing Go, or categorizing images, is all that these machines can do, unlike humans who can both understand the language, learn to perform a variety of games, and recognize images.
Despite this reality, exaggerated media reports of advances in robotics have infiltrated public debate. As a result, lawmakers could be pushed to adopt premature regulations, alert the signatories of the letter.
Roughly, the experts accuse proponents of this proposal intentionally overestimating artificial intelligence capabilities to fool an easily impressionable audience: “From a technical point of view, this project has many biases based on an overstatement of the actual capabilities of the most advanced robots, a superficial understanding of their unpredictability and their self-learning abilities and a distorted perception of the robot by science fiction as well as recent announcements of a sensationalist press,” they say.
According to some researchers, no robot has been able to convey such incorrect notions about robotics as the robot Sophia: a humanoid robot that made its first appearance in March 2016, then was granted Saudi citizenship, a title of the United Nations and inaugurated the Security Conference in Munich this year.
“I’m not against the idea of a robot for the show,” said Sharkey, who co-wrote the letter and also co-founded the Foundation of Responsible Robotics. “But when they start training at the UN and giving the nations a misconception about what robotics can do and the state of the AI right now, it’s very, very dangerous.”
He added, “This is very dangerous for lawmakers as well. They see it and believe it because they are not engineers and there is no reason for them not to believe it. ”
Granting legal personality to robots other problems of an ethical and legal point of view: for example, the robot could then claim to human rights such as the right to dignity, the freedom to integrity, the right to remuneration or the right to citizenship.
Experts also insist that this legal status of robots cannot be compared to the model of the legal person because unlike a society, a robot has no physical person behind him to represent and direct it.
In their letter, the 156 AI experts argue that current EU civil laws are sufficient to address liability issues.
Delvaux for his part said that the idea of an electronic personality is not to give robots human rights, but to ensure that a robot is and will remain a machine with a human being in control. The project is still under discussion in Brussels:” Maybe, in the end, we will conclude that this is not such a good idea,” said Delvaux.