CLIPPY IS NOT A ROBOT -- YET / BARRON´S MAGAZINE

Clippy Is Not a Robot – Yet

By Jon Swartz




Microsoft intends to put more intelligence into artificial intelligence.

Stepping up a push from a year ago, the software behemoth this week outlined plans to leverage its worldwide computing platform to gain an edge in a market often defined more by hype than commercial applications.

“We know artificial intelligence will transform every business,” Harry Shum, executive vice president of Microsoft Artificial Intelligence and Research, said at a San Francisco briefing, where AI enhancements to the Bing search engine were announced.

Microsoft (MSFT) intends to flex its AI muscles with data flowing into its franchise products and services. Bing is a distant second to Google’s search engine, but with 450 million users, it’s no slouch. There are more than 500 million active Windows 10 devices with Cortana, the company’s AI-based personal assistant. Microsoft Office 365, the dominant computing suite with 150 million monthly active users, has added AI into such functions as proofreading and automatic language translations. Its Azure cloud-based services and HoloLens, its mixed-reality smart glasses, also are using AI to greater effect.

“They are one of the largest depositories of back-room data” in the world, says Tim Bajarin, principal analyst at Creative Strategies. “How it is going to use it in commercial ways is what interests me.”

There is a catch. A sticking point for Microsoft is that two of the other companies – Google parent Alphabet (GOOGL) and Apple (AAPL) – have the advantage of widely used mobile technology to deliver AI experiences.

Through their mobile platforms, Apple (iOS) and Google (Android) collect a treasure trove of data via Siri, Google Assistant, Google Now, and other smartphone features.

Lacking a mobile piece, Microsoft has tried to fill the void: It developed Cortana for Apple’s iOS and Android, and it signed a deal with Amazon (AMZN) to link Cortana to Alexa, Amazon’s AI digital assistant.

“Mobile is a hurdle, but so it is for Amazon,” which doesn’t have a mobile platform, says Carolina Milanese, principal analyst at Creative Strategies. “Microsoft’s progress might seem incremental, but it is delivering AI through big platforms.”

AI is a natural extension of Microsoft’s legacy, extending its reach in more intuitive ways as the computer landscape blossoms into the Internet of Things, big data, cloud computing, and neural networks, according to Shum of Microsoft.

The announcement by Microsoft is the latest in a slew of consumer and enterprise deployments this year that have legitimized the market, although not all have been successful and many are narrow in scope, such as text analysis of inventory via deep learning.

AI is the rare technology that has more immediate financial upside in the consumer market than for enterprise, says Jessica Groopman, a research analyst specializing in AI at Kaleido Insights. Citing public data, she says that AI in consumer use (hardware, software, and services) should balloon to $42 billion in 2025 from $1.86 billion in 2016. She adds that the enterprise market has greater long-term growth potential, particularly as smart cities and autonomous cars gain momentum.

At the same time, a handful of recent research reports strongly indicate that after years of hype that didn’t match market realities, AI is poised to find its way into self-driving autos, image and speech recognition, and data analysis.

The economic implications are broad and sweeping: By 2030, up to a third of the American workforce – 16 million to 54 million – will need to switch to new occupations, depending on adoption of AI, according to a report published by McKinsey Global Institute in November.

From Apple’s Siri to self-driving cars, technology displayed by machines for years has been a lightning rod of debate over its long-term impact. Tech’s largest players extol its benefits in pursuit of business opportunities and futurists like Max Tegmark, president of the Future of Life Institute, insist “amplifying our human intelligence with artificial intelligence has the potential of helping civilization flourish like never before – as long as we manage to keep the technology beneficial.”

So far, that’s been a big if. Adoption of the technology has been impeded by novices – who are typically intimidated by new technology that could cost them jobs or worse -- and deep thinkers, who foresee machine learning that runs amok.

Tesla (TSLA) CEO Elon Musk famously referred to AI as “summoning the devil” and theoretical physicist Stephen Hawking warned it “could spell the end of the human race.” Not to be outdone, Microsoft co-founder Bill Gates cryptically wrote, “I am in the camp that is concerned about super intelligence.”

AI and robots, its critics whisper, will replace 7% of U.S. jobs by 2025, according to a report they cite by market researcher Forrester. Although 9% of jobs (8.9 million) will be created in fields like robot monitoring professionals, data scientists, automation specialists, and content curators, 16% in fields like office and administrative support will be lost.

It gets more sinister.

In the wrong hands, machine-learning systems powering AI could be usurped by the powerful, abetting authoritarian regimes, caution researchers.

Kate Crawford, a research scholar at Microsoft, made the case in a talk on misused data, “Dark Days: AI and the Rise of Fascism,” at the SXSW technology conference in Austin in March. From Nazi Germany’s use of Hollerith tabulating machines to isolate Jewish citizens, to private sector abuses in technology, history is littered with examples that portend potential issues with AI, she said.

Case in point: Palantir Technologies, the data-mining start-up backed by billionaire investor Peter Thiel that assisted customs officials in documenting and tracking immigrants, according to a report in The Verge , which cited public records.

It’s a warning that’s heeded by some of AI’s biggest players. Microsoft, Amazon, Apple, Facebook (FB), Google and IBM (IBM) are among members of Partnership on AI, a consortium dedicated to best practices for AI.