Could AI Innovations Be the Solution to Productivity?

Apr 9, 2017
Post Masthead

The idea of AI (artificial intelligence) as a productivity solution is nothing new. In fact, automation is something that has been talked about often as people scramble to make some sense of the “new world” of work.

For you, a knowledge worker or business owner, AI might mean new ways of doing work or changes in the tasks that you’re required to do.

From the perspective of business owners and employers, AI is bringing new opportunities to innovate and seek answers to those age-old questions of productivity. It’s not without controversy - we’ve all heard of jobs being replaced by automation and the resulting heartache for those who are unemployed as a result.

Is AI a solution to productivity goals? Let’s explore:

Current developments in AI

Whether we immediately recognize it or not, AI innovations have been creeping into everyday use for quite some time. We haven’t quite reached the capabilities of Ava, the robot from the movie Ex Machina who had human-level consciousness and social awareness, but we’ve reached a point where artificial intelligence is already quite useful.

You see it with the use of bots in apps, some of which you might already be using. At it’s basis, AI employs “machine learning” to pick out patterns and and offer choices or make predictions based on what the bot has gathered from those patterns.

For example, chatbots are being used with increased frequency among various businesses or services that need to respond to customer service queries. AI platform IBM Watson looked at how AI can enhance customer service experiences and found that roughly 50 percent of all customer service calls handled annually go unresolved, while 61 percent of those calls could have been resolved with better access to information.

Enter the chatbot. These work by Natural Language Processing (NLP) and are best suited for situations where its application is constrained. For example, messaging app Slack has Slackbot to help out with basic requests, but if it can’t help (or doesn’t understand), it will search the knowledge base or refer the user to human help.

So, in this case and the many others where a chatbot has been developed, they can quite handily reduce the load on human customer service operators so that they’re able to focus more on those customer service queries that require extra attention, rather than basic questions. Definitely a productivity improvement.

Look at Google’s “G Suite” of productivity apps and you’ll also see AI innovations at work. For example, users can access auto-generated responses where they need to reply to an email that only requires a short response.

"A year ago, Smart Reply launched, offering auto-generated replies for emails that only need a quick response. Now, more than 10 percent of all replies on mobile are sent using Smart Reply. The reception has been so strong that we're continuing to apply machine intelligence across our suite to solve customer problems.” (Prabhakar Raghavan, VP Google Cloud).

8384110298_da510e0347_z (2)

Another aspect of AI is its ability to help streamline the workflows of busy people; an example IBM gives is that of the “expert” who finds their day interrupted a lot with requests for information. Accurate knowledge transfer is important, but it can seriously impact the smooth flow of the person’s core work. AI solutions allow organizations to put in place systems where people are able to use natural language (like those chatbots) to ask their questions and receive suggested options out of a comprehensive knowledge base.

One of the biggest challenges is the lack of any form of human-like consciousness as seen by Ava in Ex Machina. It takes hundreds of hours of repetition for robots to learn basic tasks, although, granted, they seem to perform the task with consistency once learned.

Basics, such as picking up objects, is one of those challenging tasks. A robot might have learned to pick up a cup of coffee, but it doesn’t understand that it will then know how to manipulate itself to pick up a sock. Researchers at Carnegie Mellon’s Robotics Institute have been working on this, allowing robots to manipulate different objects in order to learn. It’s still something that has a way to go in terms of development though.

AI and productivity

“They’re coming for our jobs” has been a common fear voiced about AI and innovations in automation, and this fear is not without foundation. As Techcrunch found, when it comes to the loss of manufacturing jobs over the last decade or so:

“85 percent of job losses stemmed from “productivity growth” — another way of saying machines replacing human workers.”

BUT, they follow this up with:

“While there are fewer jobs, more is getting done. Manufacturing employees are better educated, better paid and producing more valuable products — including the technology that enables them to be so much more productive.”

So, what of knowledge workers? How is AI impacting productivity in the knowledge sector? It turns out companies are already using AI initiatives to replace human workers in some knowledge-based jobs, too. For example, Fokuku Mutual in Japan is using “IBM Watson Explorer” to replace 34 human insurance claim workers. They expect that Watson AI will improve productivity by 30% and allow quicker data processing, which therefore results in faster claim processing by the humans who process final payouts.

Then there’s this from The Economist:

“Bank of America Merrill Lynch predicted that by 2025 the “annual creative disruption impact” from AI could amount to $14 trillion-33 trillion, including a $9 trillion reduction in employment costs thanks to AI-enabled automation of knowledge work; cost reductions of $8 trillion in manufacturing and health care; and $2 trillion in efficiency gains from the deployment of self-driving cars and drones. The McKinsey Global Institute, a think-tank, says AI is contributing to a transformation of society “happening ten times faster and at 300 times the scale, or roughly 3,000 times the impact” of the Industrial Revolution.”

This might sound like a bit of a double-edged sword. If you’re a knowledge worker who relies on employment, you’re probably wondering what parts of your job could be taken over by AI, while if you’re a business owner, you’re possibly pondering what long-term savings and productivity gains you might get from an AI investment.

One thing that’s important to remember is that current AI is available in very narrow bands. The technology can learn via repeating a very narrow set of tasks within a given context, but it can’t replace the complete cognitive abilities of a human. As Venturebeat points out, this would be known as “General AI” and is still a long way off. A bot cannot think creatively, for example.


The future of productivity

A joint study from Goldsmith’s, University of London and IP Smith found that, as AI continues to advance, businesses will need to develop a balance of artificial and human intelligence as different roles require a mix of the two. “It is this new configuration of humans working alongside intelligent machines that will be the source of sustained competitive advantage.”

It followed by stating that by “automating and redeploying humans away from repetitive jobs to tasks that require creativity and innovation, organisations can increase productivity three times over.”

With our minds freed from those repetitive tasks, we’re able to devote more energy to the creative tasks that AI just can’t handle (at least, for the foreseeable future). Humans are expected to be needed still for 80% of complex problem solving tasks. Alongside this prediction is another: that there will be new, high-skilled jobs for humans that we haven’t even considered yet. This pattern has been proven previously where jobs have been given over to automation.

Final Thoughts

Are AI innovations a solution to productivity? Studies show that there are definite productivity gains to be made by using AI, but it is not the solution to productivity needs in business.

Today’s AI tech is very good at taking over tasks within a very narrow context and completing them to a consistent standard. For example, tasks that are repetitive and can be “machine learned” over time.

We’re not yet at Ex Machina though. Consciousness and creative thought are still solely the domain of the human operator; however, free up the human from those repetitive tasks and a more productive focus on complex problem-solving can follow.