Films depict intelligent machines as much like people, but the emerging reality is one of devices that do one task very well using artificial intelligence.
In Alan Turing’s seminal 1950 paper Computing Machinery and Intelligence, the father of modern computer science proposed a test to determine whether machines could think. If it could dupe humans into believing it was one of them, the machine passed the Turing Test.
This image of human-like generalised artificial intelligence (AI) has dominated our thinking in the almost seven decades since.
In today’s world, a different AI is creeping up on us, thanks particularly to improvements in computer-based vision and speech recognition, and to advances in natural language processing, which boosts computers’ ability to understand not just words but the meanings of casual human statements.
These devices are less human-like than the movie robots of films such as Star Wars. Yet the technological end product is still highly useful – software that’s task-specific and embedded in devices like drone navigation systems that dodge obstacles, banking chatbots that answer financial questions, and surgical robots that help doctors to complete complex medical procedures.
"Just how far can technology take up, and are we ready to go there?"
Google, one of the world’s AI leaders, is perhaps the most advanced in testing driverless cars that gather and process massive amounts of data on their surroundings.
Google is also building a new AI lab in Montreal dedicated to “deep learning”, where software attempts to mimic activity in the brain’s centre of thought formation, the neocortex. Apple recently hired its first director of AI – Russ Salakhutdinov, an associate professor in the deep learning department at Pennsylvania’s Carnegie Mellon University.
Investors are taking note, too. Goldman Sachs technology, media and telecom group co-chair George Lee told US news network MSNBC in October 2016 that artificial intelligence would “change our world ... and even the course of our species in ways that are hard to predict today”.
This current round of AI advances raises questions even tougher than Turing’s original test. Who is responsible when a driverless car crashes or an intelligent medical device fails? Just how far can technology take us, and are we ready to go there?
5 ways accountants have mastered AI
8 sectors feeling the effects of AI
An ongoing Stanford University project called the One Hundred Year Study on Artificial Intelligence (or AI100 for short) has identified eight areas where AI is already having, or is projected to have, the greatest impact. These include transport, healthcare, education, public safety and security, and employment and the workplace.
The Stanford project’s first report predicts that transport will be one of the initial domains in which people will be asked to trust an AI system for a critical task. As cars become better drivers than humans, the report forecasts that city dwellers will own fewer of them. In the US and Canada in 2030, it suggests, AI will not just be embedded in cars but in trucks and possibly even “flying vehicles and personal robots”.
In the next 15 years, the report suggests, AI can also improve healthcare by enabling the sharing of health data from personal monitoring devices like fitness bands and smartwatches .
It points out, however, that IT applications “have been slow to move from the computer science lab to the real world” and says AI has been slow to win adoption in existing healthcare systems, with both government and commercial hurdles still to be overcome. It also cautions that AI systems must win the trust of both care providers and patients.
On the education front, technologies like natural language processing platforms are expected to augment instruction by humans. Already in development, the report says, are “interactive machine tutors ... matched to students for teaching science, math, language and other disciplines”.
As in healthcare, however, AI will need to work with the existing system – in this case, teachers in classrooms.
The report underlines that one of AI’s greatest challenges over the next 15 years will be gaining public trust. However, it predicts that drones, cameras and software to analyse crime patterns may use AI in ways that reduce human bias and enhance safety without loss of liberty.
While the report states that AI will replace some jobs and create new areas of employment, it argues that in the longer term, AI may be thought of as a “radically different mechanism for wealth creation”, in which everyone should be entitled to a portion of the world’s AI-produced treasures.
“It is not too soon for social debate on how the economic fruits of AI technologies should be shared,” says the report.
CPA Q&A. Access a handpicked selection of resources each month and complete a short monthly assessment to earn CPD hours. Exclusively available to CPA Australia members.
Toward a standard of ethics for Artificial Intelligence
Experts are addressing the uncertainties of AI in other ways, too. Google, Amazon, Facebook, IBM and Microsoft teamed up in September 2016 to create the Partnership on Artificial Intelligence to Benefit People and Society. The alliance aims to set out a standard of ethics for the development of AI; non-corporate and corporate members have equal representation on the partnership’s board.
As the technology giants and other companies increasingly use AI to compete on the commercial stage, they may need to dispel public concerns about where the advances are heading.
“We’re in a golden age of machine learning and AI,” declared Amazon’s machine learning chief Ralf Herbrich when the partnership was announced. The partnership would help to “improve customer trust and benefit society”, he added.
The Silicon Valley partnership, along with studies such as AI100, demonstrates how discussion and debate are vital in determining how AI may enrich our lives. As the Stanford study notes: “Policies should be evaluated as to whether they democratically foster the development and equitable sharing of AI’s benefits, or concentrate power and benefits in the hands of a fortunate few.”
How insights and investment will impact AI
As artificial intelligence (AI) begins to drive business analysis, investment funds are flooding into the sector. US-based research firm Forrester says its survey data suggests investment in AI will triple in 2017. Software firms such as Adobe, Google, IBM and Salesforce are already building AI into their programs and will look to do more.
That should be a bonanza for a host of businesspeople looking to better understand their marketing, product management and e-commerce activity, as well as many other aspects of their business. It will also shift the balance of power in the business of consumer insights.
AI and various related technologies “will steal US$1.2 trillion per annum from their less-informed peers by 2020”, Forrester predicts in a November 2016 study.
One factor driving this change: the rising stock of devices, from mobile phones to smartwatches to cars, that now report a rich stream of intelligence about their owners’ behaviour. This new crop of pint-sized, connected computers is “flooding the market with inbound, spatially related customer data”, says the Forrester study.