Can AI solve information overload?

In some ways, artificial intelligence – in the form of automated features within popular applications – is already helping us combat info-glut. Those small steps are leading inevitably to a future in which we’ll all rely on AI for daily assistance with mundane tasks.

stressed man busy over worked overload burdened
Credit: Thinkstock

You walk into your office and click a button to login. A screen pops up showing your tasks for the day, but you notice some subtle changes. Because your task management app is connected to the computer systems in your car, it knows you’re running late. That task due before lunch is going to have to wait, because you have a meeting with the board in five minutes. 

You decide to check your email. With one click, you weed out all of the messages that aren’t important and see only the people who are important to you. No need to reply. Your email app reads the messages, syncs with your scheduling system and task list, and automatically responds using your tone, information culled from other apps, and your task load. 

In many ways, these automated features are already starting to help us combat information overload. Yet there’s an incredible amount of work still being done to help us become even more productive, offload highly repetitive tasks like answering email and tweaking a schedule, and focus on mission-critical topics and strategic decisions. 

To witness some of the best machine learning in existence, just use the Google app. When you ask a question, you can engage in a conversation. For example, if you ask “what is the weather like in New York” and then ask “What will it be like next week” the app will understand you. The Google app also lets you inquire about traffic delays, set reminders and make calls. 

Artificial Intelligence is strictly defined as a machine thinking with intelligence. Maybe having an auto-reply feature for email or a weather or traffic bot doesn’t seem like AI, but in the near future, machine learning will give us a reprieve from the digital chaos we’re all fighting against. 

Productivity software meets AI

Some of the recent advances in productivity software look like baby steps in AI programming. Yet, they point to a future when “thinking machines” will actually remove some of our daily stress and free up time for us to focus on more important things. One of the best examples of an incremental improvement is Google Inbox, which now has a Smart Reply feature. When you use a mobile device, Inbox will pop up a few automated messages. For example, if a colleague asks if you can meet later in the day, Smart Reply will give you a few options to respond yes or no. Or, if someone asks you a question, you can pick a reply like “I will find out” and click send. 

[Related: 5 apps to help you make better decisions] 

It’s a small step, but Google is using machine learning to understand what the email sender wants. In the near future, Smart Reply could handle more complex messages by noting how you typically reply to a request (say, you never want to meet before 10 a.m.) and respond for you. 

Surprisingly, another good example of AI saving us time is Evernote. If you create a new note and start entering information, you can use a feature called Context. It automatically pulls in information from other notes you’ve added. If you have a shared notebook, you’ll see related notes from other users. And, the feature pulls in notes from Web sources and makes suggestions on sites that can help. 

These examples show how AI could evolve rapidly and even help us prioritize our day. According to the experts, this is the real goal of AI for productivity: To help us complete tasks and finish projects without having to focus so much on mundane activities. 

“Imagine when software can help you prioritize what you should do and, equally as important, should not be working on,” says Alan Lepofsky, a vice president and principal analyst at Constellation Research. “What if software could look at information across a variety of tools including collaboration apps, social media, news, weather, sales pipelines, customer support tickets and more, then help you prioritize your work?” 

Decision-making 

For IT leaders, AI can also help us make decisions better and faster. As Louis Rosenberg, the founder and CEO of Unanimous A.I. explained, there are already good examples of AI in the financial sector that can watch for patterns in the stock market and automatically initiate trades. In the medical field, software can look for patterns with symptoms and genetics and make a better diagnosis. Rosenberg says this could lead to a 50 percent savings in healthcare costs. 

Of course, we are already seeing how this works with business intelligence dashboards like the one from Domo. It’s not technically AI, but the dashboards help us make better decisions because we can parse the data easier. We see a visual representation of quarterly sales or software uptime and can then react appropriately. Eventually, dashboards will include AI components that make automated decisions based on the collected data. 

“These solutions don't have to be orders of magnitude better than us at what they do to be useful,” says Stowe Boyd, a managing director for research at Gigaom. “In fact, a tool that does something no better than me but frees me from doing it – like a concierge bot that would make flight and hotel arrangements for me – would be worth a great deal.” 

Boyd says AI could eventually help with decisions in IT that are notoriously difficult for humans because of our biases and, ironically, our digital overload. One example is in evaluating the skillset for a potential hire or even promoting someone based on analytics rather than a hunch. 

There’s a real-world example of this, he says. Direct marketing company Harte-Hanks uses an AI program called Cornerstone that picks the best call-center candidates. The company found new hires answered calls 15 percent faster than those hired through the normal recruiting method. 

Of course, not everyone agrees that AI is capable of answering our email just yet. Dan Tochinni is a noted AI expert who worked at Google on the AdSense program and is now developing The Grid, which uses AI to build a Website automatically. 

“My experience with higher level decision making, even prioritizing email, is that it is a non-trivial machine learning problem which will likely take years to get right. We probably have a decade before we do, at which point deciding what to automate,” he says. 

1 2 Page 1
Shop Tech Products at Amazon