Machine learning and other variations of artificial intelligence (AI) are expected to proliferate in the enterprise in 2017. The majority of IT players, including today’s leading technology companies, have invested in the space and plan to increase efforts for the foreseeable future, according to analysts who cover the market.
However, while machine learning has generated tremendous interest throughout the enterprise, a wide gap still exists between research and beneficial use cases in the real world. And only a small number of companies have the resources to actually drive AI innovation and deliver it to the masses, sources say.
During last month’s Code Enterprise conference, LinkedIn CEO Jeff Weiner said AI was one of leading factors in the company’s decision to be acquired by Microsoft. A limited pool of AI experts means relatively few companies can make machine learning work at scale, he said at the time.
AI sets stage for big change in 2017
Box CEO Aaron Levie says machine learning and AI are pillars of his cloud-storage company’s strategy. He predicts AI will have a momentous year in the enterprise in 2017. “We are going to see a dramatic change in how enterprise software is designed and how enterprise software takes advantage of all the data that’s in our platforms to produce way better outcomes for customers,” he says.
Machine learning has already become “table stakes for data preparation and other tools related to managing curation of data,” but the technology will continue to grow and find its way into more applications and services in 2017, according to an Ovum report on tech trends in 2017. The research firm predicts machine learning is still more likely to show up in large-scale services than custom-developed apps, because few organizations outside of the Global 2000 have data scientists with the appropriate skills on their staffs.
Machine learning took the place of big data as the “shiny new thing” in technology, and it will be the “biggest disruptor for big data analytics in 2017,” according to Ovum. Enterprises are also under pressure to make data science a team sport, because the most rigorous models and hypotheses will require outside collaboration to reach greater potential.
Apple, Facebook, Google and Microsoft all open-source or share their latest research in AI to advance developments in the space. These moves from such notable organizations also meet the collective interests of scientists and researchers who prefer to share their findings with the larger community, instead of limiting access to a select group.
Apple makes secrecy exception for AI
Earlier this month, Apple made a significant exception to its generally-secretive practices and allowed its AI team to publish research papers on the subject for the first time, according to Bloomberg.com. Russ Salakhutdinov, Apple’s director of AI research who also studies the technology at Carnegie Mellon University, reportedly confirmed the change in policy at the Neural Information Processing Systems conference.
Patrick Moorhead, president and principal analyst at Moor Insights & Strategy, says he expects Apple to build out AI features for Siri with more capabilities across iOS and macOS. “If Apple does enhance desktops, I would expect it to support far-field commands like an Echo or Google Home,” he says. However, Moorhead believes a separate AI-fueled device such as Amazon’s Echo is unlikely.
Apple CEO Tim Cook has also played up the strengths and promise of AI as it pertains to the company’s future. During an interview with Nikkei Asian Review in October, he emphasized the relative immaturity of smartphones and predicted that AI will result in an “incredible future” for the iPhone.
Facebook explains foundational concepts of AI
Facebook’s head of AI research, Yann LeCun, this month produced a series of educational videos that outline how AI works, what it can conceivably achieve and how people can get involved. “AI is not magic, but we have already seen how it can make seemingly magical advances in scientific research and contribute to the everyday marvel of identifying objects in photos, recognizing speech, driving a car or translating an online post into dozens of languages,” LeCun wrote in a blog post.
The technology will be the “backbone of many of the most innovative apps and services of tomorrow,” but it remains a mystery for many people who will eventually see AI influence their daily lives, according to LeCun. “Increasingly, human intellectual activities will be performed in conjunction with intelligent machines,” he wrote. “Our intelligence is what makes us human, and AI is an extension of that quality.”
LeCun also predicted that health care services and transportation will be among the first industries that AI transforms.
“The most meaningful thing Facebook can do in AI in 2017 is to make their chatbots useful, as so far they are weak and lack slick utility,” Moorhead says. “Consumers are using them a few times, see they don’t do much well and stop using them.”
Facebook CEO Mark Zuckerberg recently published a detailed year-end update on his personal challenge to build simple AI to run his home. Zuckerberg spent about 100 hours building “Jarvis” and concluded that even if he spent another 1,000 hours on the project, he still wouldn’t be able to build a system that could learn new skills on its own. Zuckerberg’s experience also reinforced his prediction that AI systems will be more accurate at reading senses than the human nervous system within a decade. “In a way, AI is both closer and farther off than we imagine,” he wrote.
Ultimately, according to LeCun, Facebook has one goal with respect to AI, and that is to understand intelligence and build intelligent machines. “That’s not merely a technology challenge, it’s a scientific question,” he wrote. “What is intelligence and how can we reproduce it in machines? The answers to these questions will help us not just build intelligent machines, but develop keener insight into how the mysterious human mind and brain work.”
Google bakes machine learning into G Suite apps
Google invests in AI for various purposes, perhaps more than any other company today, but it also exemplifies how machine learning can be blended into popular apps, such as Gmail, that billions of people use every day. Google’s self-driving car project, now a separate company called Waymo, may generate more attention for AI, but small features in Google Apps could potentially add up to a bigger impact for more people.
When Google renamed its suite of productivity apps as “G Suite,” it also announced plans to bake more AI and machine learning into the portfolio. The company introduced a new feature in Drive, for example, called “quick access,” that “uses your activity patterns to predictively serve up to you the file that you need,” said Prabhakar Raghavan, vice president of Google Apps, in September. “Explore,” a new feature in Docs, Sheets and Slides, automatically applies formulas to company data based on common queries from users.
Moorhead doesn’t expect Google to add many more features to Google Home in 2017 and says it’s also unlikely that the device will exceed the capabilities of Amazon’s Echo in the coming year. He does, however, expect Google to extend many of its AI-driven features in image search to video search during 2017.
Microsoft shares progress on AI development
Microsoft this month shared some of the latest research and development progress it achieved in AI with a specific focus on conversational computing. The company says it has invested in AI for almost 25 years and is determined to bring new technology to consumers, business and developers in 2017. Microsoft’s newest chatbot Zo, which it launched in the United States on Kik in October, has more than 115,000 users, according to the company.
The Microsoft Bot Framework attracted 67,000 developers to date, and Microsoft says it’s adding new tools to make it easier to create bots with cognitive functions and natural-language processing. Microsoft also announced a new service, called “Calendar.help,” that uses AI to simplify the task of scheduling meetings and introduced the Cortana Devices SDK to make its digital assistant available to hardware manufacturers who want to build smarter capabilities into their devices, according to Microsoft.
“Microsoft surprised everyone in 2016 and I’m expecting more surprises in 2017,” Moorhead says. He predicts the company will release a standalone intelligent agent and a personal computer that will outperform Amazon’s Echo. “I’m also expecting to see many more business-related AI APIs.”
Gauging the impact of AI in the enterprise in 2017
Forrester Research recently surveyed 612 business and technology professionals to determine the scope of AI research in enterprise. While 58 percent of the respondents said their organizations are researching AI, only 12 percent said they use AI systems at work. “This gap reflects growing interest in AI but little actual use at this time,” the firm wrote in a separate report on the potential of AI in the enterprise in 2017.
“We expect enterprise interest in, and use of, AI to increase as software vendors roll out AI platforms and build AI capabilities into applications,” the firm predicted. “Enterprises that plan to invest in AI expect to improve customer experiences, improve products and services, and disrupt their industry with new business models.”