OpenAI is expanding its efforts to convince global governments to build more data centres and encourage greater usage of artificial intelligence in areas such as education, health and disaster preparedness.
The initiative – called OpenAI for Countries – will expand the reach of its products and help close the gap between countries with broad access to AI technology and nations that do not yet have the capacity, the company said.
OpenAI also hopes to encourage deeper usage of its tools, adding that AI systems are capable of more complex tasks than many people realise.
“Most countries are still operating far short of what today’s AI systems make possible,” the company said in a report shared with Reuters.
OpenAI started the international initiative last year and appointed former British finance minister George Osborne to oversee the project in December. Osborne and Chris Lehane, OpenAI chief global affairs officer, are pitching government officials on the project this week in Davos.
The initiative is part of a broader strategy that has helped cement ChatGPT creator OpenAI at the vanguard of the modern AI boom. The company was most recently worth $500 billion and is exploring a public offering that could be worth as much as $1 trillion.
Eleven countries have signed up for OpenAI for Countries. Each deal is structured differently.
Estonia, for example, is embedding OpenAI’s education tool, ChatGPT Edu, into secondary schools across the country. In Norway and the United Arab Emirates, OpenAI is working with other companies to build data centres and become their first customer.
On Wednesday, OpenAI executives said they were hoping to work with governments in other areas, like disaster planning. In South Korea, OpenAI is exploring a deal with the government’s water authority to build a real-time, water-disaster warning and defence system against water problems driven by climate change.
In its report, OpenAI said its typical “power user,” or those in the 95th percentile, reaches for OpenAI’s advanced reasoning capabilities seven times more often than a typical user. There are also big gaps within countries.
For example, in Singapore, which has broad access to AI tools, people send more than three times more messages about coding than average, the report said.
