Your 2019 AI Predictions
Your 2019 AI Predictions
What can we expect of AI in 2019?
Join the DZone community and get the full member experience.Join For Free
Bias comes in a variety of forms, all of them potentially damaging to the efficacy of your ML algorithm. Read how Alegion's Chief Data Scientist discusses the source of most headlines about AI failures here.
AI is nothing new, and academics have been at work with ideas like cybernetics and the brain-computer interface since mid-last century. But IoT and it's data deluge, data lakes, and advances in Machine Learning have made AI step outside of academic to mainstream enterprises and into our homes, cities, and workplaces. From healthcare to cybersecurity to manufacturing, enterprises are working hard to integrate AI into their products, devices, and services.
Artificial Intelligence has developed to achieve a wide range of capabilities from image recognition to language processing and data analysis and predictive analytics. It enables machines to make decisions and deductive reasoning faster than humans. It creates an extra layer on what IoT can achieve, almost an inter-dependence, such as in the growth of AI-powered analytics platforms for the enterprise market and the enablement of predictive and prescriptive analytics and adaptive/continuous analytics. Each year, what is possible succeeds that of the proceeding ten years.
But where there's hype, there's also plenty of fluff. One of my favorite articles of this year by The Guardian shared the practice of "pseudo-AIs” — essentially prototyping AI with human beings. The included tasks that consumers believed were being conducted by machines such as converting voicemails to text, calendar scheduling, and scenarios where chatbots were actually real people. It's a combination of companies needed to acquire enough data sets to enable to the machine to learn for themselves and take over tasks, and ambitious startups promising tech that is theoretically possible, if not quite in existence — yet. The reality is, AI is hard to do even at a basic level. I've been using an AI-embedded speech-to-text translator for the last three years that is still incredibly patchy in terms of accuracy — it's only this year that it recognizes the term IoT (it used to write coyote.)
We can expect that AI will lay the foundation for an acceleration in innovation over the next few years, boosting some sectors of the economy and completely restricting some industries. It might not be quite ready to take our jobs, but what can we expect in 2019? I talked to a number of people enmeshed in the challenges to find out:
Investments in AI Become the Norm to Remain Competitive
"The two biggest areas we're seeing as major drivers of value and growth are in applying AI techniques to retail and social interactions — although healthcare and finance represent huge opportunities. Companies in these segments have the opportunity to gather large volumes of data from their customers/users and can run their acquisition and selling funnels much more efficiently by providing a more focused product or interaction recommendations in order to keep users engaged more deeply and longer. Given the drastic paradigm shift some large enterprise players are imposing on these segments, it can be expected in 2019 that investing in such AI applications will become the norm due to a growing necessity to drive growth and retain market shares in an increasingly competitive market."
Pedro Alves Nogueira, PhD, Head of the Artificial Intelligence and Data Science Specializations and the Director of Engineering at Toptal
New Voice-Led Customer Analytics
Voice commerce to gain importance – As more and more services/websites integrate with voice capabilities like Siri, Alexa, Google Assistant, etc., organizations will need to redefine their SEO strategies to drive similar/ higher conversions. This, combined with the commoditization of voice-bot/chat-bot capabilities, should lead to newer customer experience initiatives.
Bhaskar Roy — Client Partner and Head, Customer Analytics, Fractal Analytics
The Emergence of AI Personal Security Agents
"AI engines that will play a role similar to antivirus software, helping us manage personal data exposure, protecting privacy, and shielding us from data misuses. Think about this: Google knows your most secret thoughts and fears, Facebook knows your entire family and social circle, who your secret crush is, what political views you publicly express, vs. what you think, places you have traveled to and liked. Mint knows more about your finances and spending habits more than even you do, and 23&me knows your entire genome."
Lana Klein — Managing Partner, Growth Analytics and AI Transformation (GAIT), Fractal Analytics
A Lack of Data Science Skills and Resources Leads a Push to Data Science Automation to Advance Ai and Ml Projects Within Organizations
"The pressure to achieve greater ROI from AI and ML initiatives will push more business leaders to seek innovative solutions. While substantial investments are being made into data science across many industries, the scarcity of data science skills and resources limits the advancement of AI and ML projects within organizations.
In addition, one data science team is only able to execute several projects a year given the iterative nature of the process and the manual work that goes into data preparation and feature engineering. In 2019, data science automation platforms will capture much of the mindshare. Data science automation will cover much wider areas than machine learning automation, including data preparation, feature engineering, machine learning and the production of data science pipelines. These platforms will accelerate data science, execute more business initiatives whilst maintaining the current investments and resources."
-Dr. Ryohei Fujimaki, CEO and founder, dotData
The Nexus Between Data Scientists and Business Leaders Causes Conflict
"Frustration among business leaders will continue to grow. For many companies, ownership of machine learning initiatives lies with data science teams. Despite being well versed in choosing, building and validating training algorithms and turning them into models to solve a business problem, data scientists are not familiar with what it takes to deploy and manage those models in production – an aspect that is typically owned by the operations teams. As a result, it often takes much longer than anticipated for companies to see the benefits of machine learning. This leaves business leaders unsure of when they’ll accomplish their machine learning goals, leading to mounting frustration."
-Sivan Metzger, CEO of ParallelM
Opinions expressed by DZone contributors are their own.