February 16-19 2020 – Pink Elephant: How AI is Driving Digital Transformation and Improving Operational Efficiencies Register Here

Five AI and machine learning predictions for 2017

 In Enterprise AI, IT Service Management Blog

Lao Tzu, 6th century Chinese philosopher, spoke truth that should dissuade me from what I’m about to share: “Those who have knowledge, don’t predict. Those who predict, don’t have knowledge.” Nils Bohr, Nobel Prize-winning Physicist, dogpiled: “Prediction is very difficult, especially if it’s about the future.


Alas, I am foolhardy. I swim with sharks. I poke sleeping bears. I also share the fascination of many tech blowhards with forecasting what’s ahead. Forthwith, here’s what will happen in artificial intelligence and machine learning in the enterprise in 2017.

First off, context… In the AI community, spirits are high. Most pundits are tripping over themselves to cliche last year’s biggest tech theme into oblivion with tired phrases like “…year of AI re-awakening,” and (my personal favorite mixed metaphor, I’m not making this up) “…when deep learning emerged from its cocoon and hit a home run.” Bleh.


Stories of ordinary people becoming extraordinary with AI like the Japanese cucumber farmer, Australian marine biologist, and crop-dusting drone farmers have dominated headlines, but for artificial intelligence to prove its value it must embed itself transparently in our daily work lives. It must demonstrate that the power of new technologies to increase productivity outweighs the cost of jobs that will be lost to automation.


Five predictions for 2017…


  1. Machine learning becomes the new mobile: we founded Aeroprise in 1999 before smartphones or apps existed because it was obvious that the future of computing wasn’t Clippy in a cube. By 2008, iOS and Android ushered in the era of “mobile first”. Sundar Pichai officially ended that era with his pronouncement at the unveiling of an AI-powered Google Translate in November. Google, and eventually all tech stalwarts, will henceforward be “AI first.”
  2. Public clouds make AI OAuth-simple: machine learning leaders Google, Microsoft, Amazon, Facebook, and Baidu have all released machine learning frameworks with APIs that make it as easy to add sentiment analysis or image recognition to apps as it was a few years ago to use OAuth to authenticate across platforms. API platform pioneers like Apigee and Mulesoft created large communities of API-savvy developers. Those developers will embrace new abilities to incorporate machine learning features into apps without dusting off the cover of a Linear Regression textbook.
  3. Conversation becomes the new interface: Facebook’s Messenger and WhatsApp platforms alone process more than 60 billion (!) messages daily. That’s an astonishing ten per living human per day. In the enterprise, messaging apps like Slack are creating a new culture of collaboration that is rapidly displacing stodgy, asynchronous email as our preferred mode of sharing information. Rumors of the demise of email continue to be premature… but the rise of conversational interfaces in the form of bots and chat-driven automation is real.
  4. The bots won’t take over: domain-specific AI-driven automation will mature in 2017. Cross-domain, or “generalized” AI, won’t. For example, IT issues will be addressed by bots that understand how to reset your password or fix your laptop. Just don’t expect them to order you coffee or book a flight. Expect neural networks to achieve vaunted “Artificial Generalized Intelligence” status in the next decade. At that point, machines will have the full range of human cognitive abilities. For as much progress as we’ve made, in 2017 it will continue to be the case that human toddlers will have abilities that far exceed those of machines. For example, both toddlers and machines can learn to distinguish dogs from cars… but only toddlers get excited by a wagging tail. For a great exploration of the topic, read Kurzweil’s captivating How to Create a Mind.
  5. Machine recognition of images, text, and speech will become commonplace: gone are the days when machines identifying faces or translating languages was science fiction. Expect manufacturing environments to use augmented reality to train assembly line workers, smart workspaces that adapt climate, alerts, and apps to your preferences, and legal document review to be conducted via automated text-mining. Machine learning will move beyond the realm of “how” and enter the realm of “why”.
Beyond 2017, expect many AI themes from 2016 to enter the mainstream public conscience. Those include autonomous vehicles (and the regulatory infrastructure to support them), designer genes (and the ethical maturity to make them feasible), and, eventually, singularity (the convergence of carbon and silicon-based life forms).


Until then, I encourage you to re-state, amend, or mock my predictions… while I commune with bots and swim with sharks.


Recent Posts
Contact Us

Please send us an email and we'll get back to you, asap.

Not readable? Change text. captcha txt

Start typing and press Enter to search