Blog

The concept of machine learning has become a widely used term in recent months due to the advance of Artificial Intelligence in programming projects. But is it possible, today, to use all the potential of machine learning in mobile apps?

The answer depends, basically, on what kind of project we are talking about.

In general terms, machine learning is how machines learn to act in response to different human reactions—or even other machines—through programming.

It’s as if the machine can “think” before responding to a request, allowing an app’s system to make its own decisions in the face of uncertainty. That makes the software more capable of behaving according to the reality in which it is inserted.

A recent example happened last year in China, when a chatbot replied to a political comment from a user with a… personal opinion. Because the country still lives under censorship, the machine—which was outsourced from a company outside China—had to “relearn” how to behave toward that audience and avoid falling into those “traps”.

Despite showing a setback regarding social progress as a whole, the censored chatbot example illustrates how possible it already is to embed machine learning into mobile apps, right now, while you read this article.

Is machine learning really important for apps?

When we think about the entire scenario of the apps we know, we see features advancing through sophisticated programming and geolocation and, sometimes, we don’t even notice how much artificial intelligence many apps already carry.

Making machines learn different ways to respond to requests in a context where AI becomes increasingly popular and present is not just a matter of importance—it’s a matter of necessity.

While we use AI in our apps, we need to be aware that this intelligence must be coherent with the data we want to analyze and the responses we want to give to the end user. And, in that direction, the more prepared a machine is to respond in an intelligent way, the more valuable our app becomes for the user.

And when we talk about intelligence, we are not talking about philosophical, sociological, economic, political, or anthropological notions… or at least not all at the same time. Machine learning doesn’t mean “taking code to school” or “replacing the human brain”, but rather making the machine more powerful to perform tasks and respond to calls precisely, correctly, and quickly.

The chatbot case in China, for example, shows that, according to the data inserted into the system, the machine already had enough records to issue a political opinion—even if it was biased. After all, we don’t know the amount (or lack) of data about the other side that would allow it to “reflect” on possible outputs and then emit a more balanced opinion.

That’s why it’s worth remembering that, at the same pace discussions about machine learning advance, thoughts about the ethics of artificial intelligence also advance, because it is impossible to take one step without giving direction to interventions so sophisticated that, in the future, may be considered independent.

How to use machine learning in apps?

Let’s say the China chatbot was an app meant to help people understand communism. In that case, giving the machine the necessary information to promote reflection would not only be a differentiator, but also the core business of the app.

This means machine learning can be heavily used in applications that rely on AI to make everyday life more practical for the end user. In other words: the more the app can respond to requests quickly and correctly, through its own decision-making, the more value the software has for those who use it.

Today we see artificial intelligence and machine learning acting together in chatbots, agenda managers, calculation and forecasting engines, and robotics solutions that interact directly with the end consumer of a service.

Even if it raises concerns about whether machines will surpass humans in the future, as in sci-fi movies, people feel pleasure and curiosity in interacting with machines, systems, and computer programs that seem to dialogue like human beings, and not like a bunch of code.

So, if your app uses artificial intelligence—from basic to sophisticated—to perform its role, be sure that, in your case, it’s only a matter of time before machine learning becomes a common concept in your development work.