- Globally, about 14% of all companies are said to be using artificial intelligence.
- Worldwide spending on AI systems is expected to hit almost $100 billion annually by 2023, according to industry researchers.
- Despite being a tech buzzword for decades, there's still some confusion about which tech qualifies as true artificial intelligence.
Every time someone asks a smart fridge to remind them to pick up milk, they're helping lay the groundwork for what's expected to be a $100 billion world-changing industry.
A fridge's artificial intelligence (AI) digests the user's words, running them through machine-learning software, which gets smarter every time you talk to it, said Dorando Doo, senior vice president at iFlytek, a Chinese AI firm. As the fridge's software gets smarter, it'll help inform a wide range of other AI systems.
"That learning capability can be applied to various industries, such as education or health care," she said at CNBC's East Tech West conference in the Nansha district of Guangzhou, China, this week.
AI is being hailed as a revolutionary technology, which could eventually improve everything from banking to transportation. It promises software that, once deployed, can continue learning and evolving, perhaps solving complex problems outside the scope of human abilities. But current AI tech is used in more pedestrian ways, with the lion's share in customer service, sales processes, and threat detection, according to research firm IDC.
Worldwide, some 14% of companies have already deployed some form of it in their processes or products, up from 4% a year ago, according to Gartner. In China, 14% of firms have embraced AI tech, an IBM executive said this week.
Businesses are expected to spend as much as $97.9 billion annually on AI projects by 2023, up from just $37.5 billion this year, IDC forecasts.
But there's still some debate about what qualifies as true AI. When China's state-run Xinhua News Agency last year announced that it had created the "world's first" AI news anchor, some in the industry balked. The digital anchor was built to move and speak as a normal anchor would, but its underlying technology leaned closer to machine learning than true artificial intelligence.
Perhaps the simplest definition of what constitutes AI came from one of the founders of the discipline, Stanford University Professor John McCarthy, who died in 2011. He defined AI as "using computers to understand human intelligence."
Even with some uncertainty about what fits into the AI discipline, businesses around the world are touting their AI software. At China's largest entertainment streaming site, iQiyi, for example, it's being used to restore old movies. A two-hour film restoration used to take some 20 days, even with 10 people working on it, said Wang Xuepu, vice president at iQiyi.
"But with this new technology, we only need 12 hours to realize this," he told CNBC's Qian Chen in Mandarin during a fireside chat on Tuesday.
Everything from video recommendations to package deliveries are now being optimized by AI software, said Song Zhang, managing director of ThoughtWorks China, a tech consulting firm. The next logical step would be to use such software to solve important social problems, like matching rural patients with doctors in cities, he added.
"We don't have to focus too much on the word 'AI,'" he said in Mandarin. "We need to be more solution-focused."
But AI is still far away from "real" intelligence as we understand it, said Rong Luo, CFO at China's TAL Education Group. His company is now investing about 5% of its resources in developing AI tech, which includes short online classes built by AI software. But it also still has live teachers doing much of the on-camera work, including looking at student responses for accuracy, he added.
"No matter how you dissect AI into different spaces, what we have is still very basic," he said during a discussion about the "Future of AI."
Industry insiders are also aware that their efforts have been seen as concerning by people worried about privacy. Others are worried about robots revolting, attempting to overthrow our societies, said Zhang.
But there's a wide gap between what the current AI industry is capable of and the public's "fear," said Danil Kerimi, head of the Fourth Industrial Revolution Network at the World Economic Forum.
"We have to find a balanced way to discuss that," he added. "Again, it's what we make of it."