More than a thousand industry experts, including Elon Musk and Steve Wozniak, have urged a six-month pause on training language models more powerful than OpenAI’s GPT-4.
The Future of Life Institute’s open letter, signed by scientists, entrepreneurs and independent researchers, urges a six-month pause on training language models more powerful than OpenAI’s GPT-4. Experts say that the development of advanced AI should be paused until universal safety protocols for such systems exist.
“Powerful AI systems should be developed only when we are confident that their effects will be positive and the risks manageable,” the statement said.
The letter outlines potential threats to society and civilization from competitive AI systems. They include economic and political upheavals.
They also urged developers to cooperate with regulators.
“The letter isn’t perfect, but the spirit is right: we need to slow down until we better understand the ramifications,” said Gary Marcus, an honorary professor at New York University and a signer of the letter.
The scientist warns that large language models could harm many people.
“Large players are becoming increasingly opaque about what they do, which makes it harder to shield society from any harm that may materialize,” Marcus added.
Industry concern has grown amid the rapid development of large language models and chatbots built on them. Since the announcement of ChatGPT in November 2022, tech giants and startups have entered a new AI race.
In February, Microsoft and Google, one after another, unveiled search services based on conversational algorithms.
In March, OpenAI opened access to the API for ChatGPT, which allowed dozens of companies to integrate the chatbot into their own applications.
Over the past two months, Microsoft has rolled out a range of AI features based on OpenAI’s technology across its products, including Word, Excel, PowerPoint, Outlook, Teams, GitHub, Bing, Dynamics and Azure.
In March, Europol warned about threats from ChatGPT.
In the same month, OpenAI CEO Sam Altman acknowledged the risks associated with the chatbot.
In January, the developers of ChatGPT warned about the use of AI by disinformation agents.