An adviser to British Prime Minister Rishi Sunak has warned that artificial intelligence systems are on track to become powerful enough to 'kill a lot of people' in just two years. The PM's adviser is Matt Clifford, who is leading the government's Foundation Model Taskforce.

AI needs to be controlled globally

Matt Clifford focuses on testing AI language models such as ChatGPT and Google Bard. He is also the Chairman of Advanced Research and Invention Agency. During an interview, Clifford said that AI has the ability to create cyber and bioweapons that can cause many deaths. He warned that AI needs to be Controlled Globally and if this is not done it could become an extremely powerful system that people may struggle to control.

There are many different types of risks with AI

Clifford said they feel there are many different types of risks with AI, and people often talk about near-term and long-term risks. Near-term risks are truly frightening. Today AI can be used to create bioweapons or carry out large scale cyber attacks. This is the worst thing. Clifford's comments came in response to a letter written by dozens of experts. He is asked what percentage of humans could be wiped out by AI, to which Clifford says that it cannot be zero. If we try to create an artificial intelligence that is more intelligent than humans and we don't know how to control it. Currently it is going to create the possibility of all kinds of dangers at this moment and in the future. However, he also said that AI can be good if used properly.