Singapore’s Vision for AI Safety Bridges the US-China Divide

Government Singapore has been released A blueprint Today is for global cooperation Artificial intellect Protection after a meeting of AI researchers in the United States, China and Europe. The document gives a shared view of working in AI protection through international cooperation instead of competition.

“One of the few countries on the planet Singapore that gets better in both the East and the West,” said MIT’s scientist Max Tagmark, who helped to convene the AI ​​Lumminar meeting last month. “They know that they will not build [artificial general intelligence] Themselves – they will do it with them – so the countries that are about to create it are very much in their interest to talk to each other. “

Countries thought that AGI was most likely to be created, of course the United States and China – and yet these countries seem to be more intended to exceed each other than to work together. In January, after the Chinese Startup DEPSEC Reveal the model of a cutting edgePresident Trump has called it “a awakening call for our industry” and said the United States “need to be laser-centric to compete to win.”

Global AI Protection Research Priority Singapore Singapore Sens Camei Researchers have called for cooperating in three main cases: studying risks raised by border AI models, exploring the safest ways to create those models and developing methods to control the behavior of the most advanced AI systems.

This Sens Came was created at a meeting held on April 2 April, along with the Learning Presentation (ICLR) on a premiere AI event in Singapore this year.

Researchers from OpenAI, Ethnic, Google Dipmind, Jai and Meta participated in the AI ​​protection event, such as MIT, Stanford, Singua and the Chinese Academy of Sciences. Experts from the United States, the United Kingdom, France, Canada, China, Japan and Korea also participated.

“In the era of geopolitical division, this extensive synthesis of AI safety-edge research is a promising symptom that the world community is united with the commitment of a partnership for the formation of a safe AI future,” Din Jeu Lan of Singua University said in a statement.

Development of growing AI models, which have some amazing skills, researchers have become worried about various risks. Some of the issues with the problem concentrated toward the near-mayor damage Biased AI system Or possibility Criminals to use the technologyA significant number of believes that AI can be a threat to humanity as it can start to surpass people across more domains. These researchers, sometimes known as “AI Dumors”, are anxious that models can cheat and manipulate people to follow their own goals.

The potential of AI has also discussed the racing of a weapon between the United States, China and other powerful countries. Technology is seen as criticized for economic prosperity and military domination in policy circles, and many governments have tried to share their own views and regulations by managing how it should be developed.