Britain announced that it will hold the first global summit on artificial intelligence security to ensure that AI benefits mankind.
CTOnews.com, June 8 (Xinhua)-- the British government announced that the first-ever global artificial intelligence security summit will be held this autumn. The summit will invite key countries, leading technology companies and researchers to discuss security measures to assess and monitor the most significant risks posed by artificial intelligence.
Rishi Sunak, UK prime minister of Tuyuan Pexels, said: "artificial intelligence has incredible potential to make our lives better, but we need to make sure that its development and use is safe and reliable."
Sunak added: "throughout history, we have repeatedly invented new technologies that subvert the paradigm and use them for the benefit of mankind, which we must do again. No country can do this alone, which requires global efforts. But with our extensive expertise and commitment to an open and democratic international system, Britain will stand with our allies and lead the way."
Artificial intelligence brings a variety of risks, and the most common danger is that artificial intelligence may produce misinformation, which is called "hallucination". OpenAI is working on a way to perform logical verification at each stage to reduce the occurrence of hallucinations.
In addition to this problem, artificial intelligence may also be used by hackers to generate malicious code for network attacks. In addition, it is also a situation that politicians must deal with when artificial intelligence replaces jobs.
CTOnews.com noted that before the summit was announced, Sunak's artificial intelligence consultant Matt Clifford said that if left unregulated, artificial intelligence could become very powerful within two years, making it difficult for humans to control. Although Terminator artificial intelligence now looks like science fiction and may remain so within two years, it is always a good thing for the government to study the potential risks before it is too late.
