Are "Mind Reading Apps" the Future?

Can an app really read your mind?

“What do you want?” It could simultaneously be one of the easiest and most difficult questions. This is because the answer spans a baffling array of possibilities from immediate physical needs, to medium-term goals, right through to long-term ambitions. But what if apps could - like a friend - get to the heart of them all?

This might sound like science fiction, but in its basic form getting to know someone in real life is actually centred on a series of interactions and remembered patterns. Your longest-running friends are often the people who ‘know you’ the best by dint of the fact they’ve seen you in a wider variety of situations. This means they more likely know how you will react at any given time.

The whole process starts small. Maybe your colleague finds out you are constantly on the hunt for unusual types of rice cakes, spots some and buys them for you. Perhaps as you get to know each other better, they discover your longer-term goal from all this rice cake munching is to lose weight. And as you become friends they finally learn that you’re back in the dating game after a lengthy, acrimonious divorce.

This represents three levels of want expressed through a variety of actions. But your real friends will understand all of them and will be ever-ready to help. This could come in the form of purchasing you exciting new snacks, suggesting exercise classes that look fun, or introducing you to people they know you might like.

If apps had all the information your friends had they would be ‘mind readers’ and it looks like we’re already on the way there. “Apps are blind and you have to make them see,” explains Mike Gualtieri of Forrester Research. “That is the first thing you have to do and that is what sensors are about. You have to have the sensory inputs.” This provides the information the app needs to learn more about your preference.

“Once it has that data, Artificial Intelligence can be more valuable, because now it can contextualise data and it can start to predict what your intent is in the moment,” he continues. “And that is what it really means to read someone’s mind: what do they want right now? What is their intent? If you know their intent then you can predict how to fulfil that intent.”

“The ‘mind reading’ app is all about sensory input, knowing what is happening in that user’s life but then also learning about the user so you can individualise it. You’re not mind reading a generic human’s mind. You’re mind reading an individual. That app has to learn about you the way a friend has to learn about you. A friend is very good at predicting your mood and your intent and what you need at the moment.”

Gualtieri believes we’re at the beginning of a journey which began 30 years ago but never had the tools to take off. “It started in the 80s with Artificial Intelligence but it failed because we were lacking three things that we now have.” These were ubiquitous computing power, a lot of data and almost always on connectivity.

“We are progressing from a mobile world where everything took [a] long [time] to be done, to a world where 1-2 clicks are enough,” says Gil Bouhnick, Vice President of Mobility at ClickSoftware, a company that builds predictive apps for field service organisations. “The future will be hands-free and mobile apps will certainly be performing more main activities for the users without any interaction. Context aware apps are here already, and they will become better as we move forward.”

Gualtieri believes two things are holding us back at present. “Firstly, the technology has to be easier to use. If you look at IBM’s Watson project they’re trying to take this very complex technology that was created by computer scientists and they’re trying to productise it.  The technology needs to improve.”

“The second thing is imagination. You need a whole new design paradigm for designing applications. Applications are designed as menus and buttons; nobody really thinks what if the app could read my mind. Designers don’t know what is and what is not possible at this time.”

“There has to be a revolution in how designers think about how they design apps. [This is] partially designers not knowing what is possible with predictive technology so they don’t even ask the ‘what if’ questions.  ‘What if this app knew I needed the toilet?’ or ‘what if this app knew I was angry?’ those would be considered science fiction-type questions, so they’re often not asked.”

Bouhnick stresses that “at the moment, context aware apps can act based on location, time, and data changes. [But] an app that is making ‘mistakes’ is much worse than an app without any intelligence, and this is one of the reasons why we need to be cautious. An app can ‘propose’ to do something, but an app that is actually making decisions for the user is very risky.”

“Once we solve that part,” he continues, “we will end up seeing apps taking more control over the actions, therefore saving user steps. [And] in that respect, enterprise apps have a potential to include more and more artificial intelligence because enterprise policies and procedures are sometimes easier to compute than individual behaviour.”  

“Every app, every website can benefit from being able to predict something with certain probabilities,” adds Gualtieri. “When we say prediction we always mean probabilities, so where it is okay to have a probability, we can use it. If it is an absolute decision you have to make that is a rule. There are some things that are life threatening and it is not good enough to predict,” but outside of that the possibilities are endless.

Gualtieri believes customer engagement is the biggest opportunity as “most people have different needs and different desires so the ability to predict is very good in customer engagement.” However, he adds: “the tip of this sphere is the internet giants [and] how they’re running big revenues from advertising. They want to stick adverts in front of you that can totally read your mind. A lot of these ads are terrible, they are not that specific [but] that is the Holy Grail.”

The flip side of this point was raised in a comment at the bottom of an article in the MIT Technology Review published last year. “Sounds to me more like ‘prescriptive’ apps,” wrote Peter P in response to the piece titled: “With Personal Data, Predictive Apps Stay a Step Ahead “which, the annoyance aside, could easily be biased towards advertisers that will pay more to get prescribed first.”

The more connected we are to sensors, the more possible all this will become. One trend which will help push this forward is wearable tech. At a recent Cisco event on the Internet of Things for example, Phil Smith, CEO of UK and Ireland, asked those with wearable devices to raise their hands. Not surprisingly only a couple of people did so, but as he pointed out: “If I ask the same question next year, I guarantee half the room will respond.” Just think, if this prediction proves correct, large swathes of the population could be beaming out all kinds of semi-medial information, direct from their bodies, very shortly indeed.

‘Mind reading’ may all sound extremely futuristic, but when you break it down to predicting behaviour based on what has gone before, it all becomes entirely conceivable. The really scary part probably is just how much data is available and just how closely it might be able to mimic the actions of a friend… who really does ‘know’ you.


Kathryn Cave is Editor at IDG Connect