Current models of artificial intelligence (AI) are “pretty stupid,” said Nick Clegg, Meta’s president of global affairs. Clegg was the Deputy Prime Minister of Great Britain in the period 2010-2015. He then accepted the job offer at Meta.
“The hype has somewhat outpaced the technologyClegg added, quoted by the BBC.
The big language models that power chatbots like ChatGPT basically connect the dots in huge arrays of text and guess the next word in a sequence. Essentially, to date, artificial intelligence is dependent on vast arrays of information and cannot “think” like a human. Its activity is rather statistical.
“Existential threat warnings issued by some AI experts refer to systems that do not yet existClegg told the BBC’s Today programme.
According to the expert, AI risks are more related to hate speech and its use by detractors.
Meanwhile, Meta and Microsoft will partner on the launch of the commercial chatbot Llama 2. It will be available and usable through Microsoft platforms such as Azure.
“I feel I can state without any fear of controversy that the LLMs we offer open source are safer than any other AI LLMs that are open source” said the former British deputy prime minister. He added that concerns about open source AI are unfounded. Clegg also criticized critics, hyperbolizing the danger from artificial intelligence.
“Meta’s open source system cannot generate images or provide bioweapon building tipsClegg said.
2023-07-19 11:03:39
#Meta #Boss #pretty #stupid
