The intelligence business looks set to become a US$40 trillion ($54.12 trillion) business, says Eric Schmidt, former head of Google and co-founder of philanthropic venture, Schmidt Futures. He was speaking as part of the keynote at Databricks’ data and artificial intelligence (AI) summit in San Francisco this year.
While there is huge potential in the industry, with organisations needing to use AI in their business models, Schmidt acknowledges that there are a few clear short-term threats that need to be addressed.
Misinformation and disinformation is one of them, especially when generative AI is going to be used widely for information, he explains.
Safety in AI is another, with these systems being able to perform “various forms of cyberattacks”.
“There's always this question of how do you test for safety issues that you don't know yet, without deploying [AI systems]? There's a general fear that when these systems are launched, they will not be fully tested, [so] they'll have some emergent behaviour that we have not yet seen and we can’t test them,” he warns.
“There are scenarios of extreme risk. And these systems are going to get regulated around extreme risks… [and] I'm talking about thousands and thousands of people harmed and killed from something,” he stresses.
See also: Australia’s social media ban for under 16s to become law
For now, Schmidt sees great potential in AI, with these systems enabling people to be more efficient in their work. “We don't know yet what the value of intelligence is. We don't know how to price it. But it's obviously high. Inference, which is the ability to actually answer the question, looks like it's going to become incredibly expensive,” he says.
The cost of training AI, however, is going to be extremely high and will go even higher, he adds. “There are people who believe that the frontier models, which cost [around] US$100 million to train today will go to a billion [US] dollars [in the future.] So that is a massive change from a software perspective. In my career, we've never seen that kind of price increase [for] software.”
