A lot of AI scaremongering centres on military artificial intelligence out of control a la Terminator, but IDC has a more meat-and-potatoes warning about bots taking white collar jobs. Photo / Getty.
US President Donald Trump plans to sign an executive order directing federal agencies to prioritise research and development in artificial intelligence.
The plan, called the American AI Initiative, comes after China and other countries have pledged big investments to advance and apply AI technology in fields ranging from warfighting to health care.
The White House plan doesn't include any funding details. The administration says it's up to Congress to appropriate money.
The order directs federal agencies to make government data and computing resources more available to artificial intelligence experts while maintaining security and confidentiality.
It also says federal agencies will establish guidance to ensure the new technologies are developed in a safe, trustworthy way.
Here, IDC research director Louise Francis says that by 2023, 40 per cent of NZ workers will be working with bots or some other form of AI.
And a report by economic consultancy Infometrics estimates that 31 per cent of current jobs in the Kiwi workforce could be automated in the next 20 years. Manufacturing is getting more and more high-tech, but software bots will increasingly invade white collar professions, such as law.
Crown agency Callaghan Innovation says AI technologies such as machine learning and natural language processing could potentially create more jobs than they destroy as they help existing Kiwi business become more competitive, and startups based around those technologies emerge, such as FaceMe, Soul Machines and Straker Translations.
But so far, New Zealand is behind other countries in encouraging AI, and creating an ethical framework, Callaghan boss Victoria Crone says.
Last year, the government moved to close the artificial gap by formalising a relationship with the AI Forum, an industry group that works with Otago University's NZ Law Foundation Centre for Law and Policy in Emerging Technologies. So far it's been big on reports and charters and light on concrete action. It doesn't help that a parallel initiative - the effort to appoint a government chief technology officer - collapsed in a heap.
An audit of all algorithms used by NZ government departments and agencies is underway, but so far there's no answer to calls for a data-mining watchdog.
National strategy wanted for US
Tech leaders have pushed the Trump administration to develop a national AI strategy. The White House in December hosted a listening session with the CEOs of Google, Microsoft, IBM, Oracle and Qualcomm to field ideas for securing American dominance in AI and other fields such as quantum computing and faster wireless technology known as 5G.
Trump made brief mention of technology at this month's State of the Union address, pledging "investments in the cutting-edge industries of the future."
US tech companies including Amazon, Google and Microsoft have promoted the potential of AI, but also warned of its perils.
Tesla's Elon Musk has been at the extreme end of the warning scale, painting the nightmarish scenario that humans could be forced to serve an immortal robot dictator.
At a more meat-and-potatoes level, there have been fears the Trump administration has failed to appreciate the degree to which manufacturing is being automated - something that was on painful display with the deal for Foxconn to open a factory in Wisconsin as billions in corporate welfare flowed to the Chinese giant with no guarantee of jobs in return.
Kate Crawford, a co-director of New York University's AI Now Institute for studying the social implications of artificial intelligence, said the directive takes some steps in the right direction but is too light on details.
"AI policy isn't an autonomous vehicle," Crawford said. "You basically need a detailed plan or it's going to run off the road."
Crawford said she welcomed the Trump administration's intention to accelerate research and regulate AI across different industrial sectors. But she said the administration also must ensure that AI's potential ethical challenges are taken seriously.
AI-based technologies such as facial recognition can be used to enhance government surveillance, while studies have found that computers are susceptible to the same racial and gender biases as the humans whose data they learn from.
Economists have also warned that AI advances could displace many US workers in the coming years — something that Trump's plan doesn't do enough to address, said Bradford Newman, an attorney who is pushing for a new regulatory body to govern AI issues.
"We can figure out how to regulate and account for the downside risks now, or we can wait until it's too late and it's purely reactive and people are out of work," Newman said.