Proposal Seeks Kill Switch And Mandatory Registration For 'Smart' Robots
The distant future once detailed and imagined in science fiction movies and books is now upon us, and humans are scrambling to stay ahead. Robots are no longer a foreign concept to the average consumer, and indeed you may even find versions of them, such as the Roomba, in the average home. Innovations in robotics will no doubt move quickly over the coming years, and will happen in step with advancements in artificial intelligence, painting a picture of a future where advanced robots can learn and think for themselves. To keep such advanced technologies in check, humans need a plan.
A new report from the European Parliament details some of the issues, concerns, and areas of focus surrounding so-called 'smart' robots, ones equipped with the AI and hardware necessary to perceive, learn about, and interact within our environment.
"Whereas now that humankind stands on the threshold of an era when ever more sophisticated robots, bots, androids and other manifestations of artificial intelligence ("AI") seem poised to unleash a new industrial revolution," the report states, "...it is vitally important for the legislature to consider all its implications."
The report lists a huge array of concerns and points of consternation, including things like 'soft impacts on human dignity' and the human ability to control its own creation should robots ever become more intelligent and capable than humans.
While those topics have long been parsed by popular media, the report also highlights potential issues less often stated: things like intelligent machines being able to pick through your personal data and share it with other artificial intelligences without human involvement and, perhaps, human awareness of the privacy violations.
The new proposal details possible steps humans can take toward mitigating these issues. One such suggestion is the creation of a registration system for "smart robots." This registration system will be used for traceability, the report says, and will be managed by the EU Agency for Robotics and Artificial Intelligence.
In addition to a system to register and track the robots, the proposal also seeks a kill switch implementation "that should be consistent with reasonable design objectives." Such a kill switch would enable human operators to terminate a robot of one sort or another quickly and, likely, remotely if the need should arise. The proposal also recommends that robots' "decision-making steps" can be traced and altered, that such machines should be made with operation that follows all applicable laws and ethical principles, and more.
One particularly interesting suggestion is the creation of tracing tools that will monitor any particular's design period, making it possible to both explain and account for a robot's behaviors. Just as well, the legal document suggests that designers need to make sure a human can tell when they're interacting with or talking to a human through some sort of identifier.
You can read the full report here.