A historic agreement between big techs and global governments is prioritizing the responsible use of Artificial Intelligence (IA). Microsoft, Amazon and OpenAI were in favor of creating the so-called “panic button”. The purpose of the resource is to establish safe AI development practices at advanced levels, minimizing problems and stopping risky actions.
In this way, the technology, also known as STOP button, would stop smart commands in suspicious and inappropriate cases, when not properly mitigated by generative solutions and programs. For that, frameworks must be created by AI developers, facilitating the identification and prevention of challenges in this area.
It will be through these frameworks that the “red lines” will be configured, valid for AI use risks categorized as intolerable. Automated cyber attacks and the production of biological weapons are some of these examples, in addition to the misuse and fraudulent use of personal data.
Pact includes accountability
This important international agreement was signed during the AI Security Summit, event held in Seoul, in South Korea. Creating the button, nonetheless, was just one of the topics discussed during the occasion. In addition to the Asian host country, the agreement was supported by the United States, China, United Kingdom, France, UAE and Canada.
In a statement, the Prime Minister of the United Kingdom, Rishi Sunak, claimed that this is the first time that leaders from different parts of the world agree on the same AI security principles. The signed pact also highlighted the need for transparency in AI processes and accountability for companies that make money from the use of technology.
Agreement without the force of law
At the moment, the generative AI market is among the fastest growing areas worldwide. The segment is expected to grow approximately 35% until 2025, symbolizing a market value of US$ 300 billion. The creation of the “stop” button is not exactly new, as technology leaders had been debating the issue for a few years.
The agreement signed, although, does not have a legislative character. Big techs committed to developing the security feature voluntarily. This is a collective initiative that will depend on individual actions by companies. Therefore, Big techs are not subject to any type of fine or punishment in case of non-compliance with the agreement.
