Greg Stanton, founder of Microsoft, stated that merely having a "kill switch" for artificial intelligence cannot prevent the misuse of this technology for attacks. In an interview aired on Sunday on NBC's "Meet the Press," he said, "The existence of a kill switch is a strange topic and is not sufficient on its own."
Serious Risks of Misusing Artificial Intelligence
Stanton continued that we have not yet reached a stage where artificial intelligence can automatically control computers and disappear. He also noted that the public discussion on this topic reflects the lack of technical awareness among many individuals.
Read more: Andy Burnham Faces New Challenges in Artificial Intelligence Program
While Stanton is not against the kill switch, he emphasized that companies must oversee advanced artificial intelligence and maintain accurate records of activities to prevent potential abuses such as cyberattacks or biological terrorism. He added, "The United States must ensure that these measures are taken. China will also want this to happen."
The Need for Legislation and Oversight
Stanton's remarks came after Senator Rand Paul (Republican from Kentucky) blocked Senator John Neely Kennedy's (Republican) attempt to quickly pass legislation requiring artificial intelligence companies to create a kill switch. Paul stated that senators need more information before implementing laws at the economic level and attempted to add to Kennedy's proposed bill the establishment of a group that could provide safety policy recommendations for artificial intelligence to the Senate Commerce Committee. This proposal was rejected by Senator Kennedy.
Stanton also called for the passage of legislation for safety and oversight requirements, stating that voluntary measures from artificial intelligence companies are not enough. He said, "No one thinks self-regulation is sufficient." In his view, the more immediate risks come from individuals misusing artificial intelligence rather than these systems acting independently. He added, "The immediate issue concerns individuals using artificial intelligence to cause harm, not that artificial intelligence operates autonomously."
Read more: Appeals Court Rules in Favor of Trump Administration on Anthropic Blacklist · Cardinal Michael Czerny Did Not Endorse the Use of Artificial Intelligence
