In a stark warning reminiscent of the nuclear arms race era, a government-commissioned report has raised concerns that the rapid development of artificial intelligence poses a significant threat to humanity.
Obtained by TIME Magazine, the report, titled “An Action Plan to Increase the Safety and Security of Advanced AI,” emphasises the critical need for immediate intervention by the U.S. government in light of the escalating risks associated with advanced AI and artificial general intelligence.
Issued by Gladstone AI Inc., the report highlights that the proliferation of advanced AI capabilities mirrors the destabilising impact witnessed during the introduction of nuclear weapons.
It underlines the pressing requirement for the U.S. government to address the escalating national security risks arising from the unchecked expansion of AI technologies.
Widespread report
The comprehensive report, developed over 13 months, engaged with over two hundred stakeholders, including representatives from the U.S. and Canadian governments, major cloud providers, AI safety organizations, as well as security and computing experts.
The proposed intervention plan outlined in the report advocates for the establishment of interim advanced AI safeguards, followed by their formalization into legislative frameworks.
These safeguards would subsequently be extended to an international level to ensure global security.
Among the suggested measures are the creation of a dedicated AI agency tasked with regulating the computing power levels at which AI operates.
Inner workings
Additionally, AI companies would be required to obtain government approval before deploying new models above specific thresholds.
The report also contemplates the possibility of restricting the dissemination of information detailing the inner workings of powerful AI models, potentially through the imposition of regulations on open-source licensing.
The report urges tighter controls on the manufacturing and export of AI chips, acknowledging the pivotal role played by hardware in the development and deployment of AI technologies.