Overnight tech bosses met with US President Joe Biden at the White House to discuss the threat — and opportunities — of Artificial Intelligence.
But the concept of a giant computer mind capable of making decisions is not a distant prospect — it's here now.
White House press secretary Karine Jean-Pierre told reporters today the president had been "briefed" about ChatGPT, a text AI that has thrown the technology into the front of the global consciousness.
The president also spoke to AI bosses at the meeting, she said.
"AI is one of the most powerful technologies of our time... we want the products to be safe."
But to understand AI, experts say it's important to understand how it was developed.
University of Auckland computer scientist Michael Witbrock has been working in this field for decades.
In the 1980s, in Dunedin, he worked with Peter Toma, who developed one of the first Machine Translation computers at Georgetown University in the USA.
These types of devices were used by the American military to automatically translate Russian documents into English and was one of the early examples of a computer having to analyse and generate text.
"I don't think it came at all as a surprise when a computer showed a decade ago it was better at playing chess than us," Witbrock told 1News.
"But I think it comes as quite a surprise that they've become good at writing poetry," he said.
"ChatGPT — it has done the equivalent of sitting down and reading for 20,000 years.
"Imagine what that can do."
He said the serious idea of an artificial intelligence that could replicate a human ability, or go beyond them, started with Alan Turing, an English computer scientist. One of Turing's most famous accomplishments included developing a machine during World War II that cracked the German enigma machine.
After the war, and into the 1950s, there were a series of conferences where the name 'artificial intelligence' was coined.
"I think one of the reasons why people are worried is that what we thought were some of the special qualities of being human, these creative things like writing stories or writing art... they've turned out to be inside the capability of computers," he said.
"They are surprisingly good for systems that were expected not to be creative at all."
Text and image based artificial intelligence systems were already making high-quaility creative content — that almost anyone could access and use.
"We should realise that something is going on and we have a powerful new technology that is going to get very much more powerful, very quickly and this change will have to be incorporated into human civilisation," Witbrock said.
"It's not going to be entirely easy, but probably can be enormously beneficial if we get it right."
Psychologist and AI Commentator Paul Duignan was running workshops for groups and organisations in New Zealand on artificial intelligence and its impact on Aotearoa.
"I did have one person, almost with an overwhelmed reaction, say 'what are we going to do about this?'," he said.
"The next step is [figuring out] how are we are going manage the introduction of this in a responsible way."
He said it was an enormous challenge for governments across the world to develop effective regulatory tools — and do it before it was too late.
"[Governments] don't really like moving really fast but unfortunately with this one they will have to move really fast, so it's really good that President Biden is starting to talk to people," he said, reference the meeting overnight.
"Governments need to move onto a faster planning mode about this, otherwise we won't be able to close the stable door after the horse had bolted — like we did on social media, it took about 10 years to even start thinking about regulating social media," he said.
"In 10 years time if we let AI run with it we are going to be in an entirely different world."
Duignan said this transformation would in some ways be bigger than the creation of the internet.
"Some people are worried about AI taking over, and that is a legitimate worry... but the earlier concern is bad actors using AI for nefarious purposes," he said.
"We're already seeing people mimic people voices... we're hearing cases of people producing fictitious material, pictures being doctored, people using it for bad purposes."
He said we are going to be "swamped with a tsunami of material" that people would not be able to figure out if it was true or false.
Despite the risks, he said, there were also big opportunities for humanity — and a "total game changer" for sectors such as medicine, technology, and education.
"There are some aspects of them where they may be able to help us with some long standing issues."
Duignan gave education as one example.
"All of sudden we have this super intelligent tutor that can tailor the information in any format that can work for a particular group.
"For example, the other day I asked ChatGPT to explain quantum physics to a five-year-old, and it started writing metaphors about waves on the beach, which a five-year-old can relate to.
"There are big opportunities to use it to address inequality."
SHARE ME