How AI Can be Used to Predict Mass Agitations

by July 28, 2017 0 comments

By Tarun Wig, Co-founder, Innefu Labs

Let’s just put aside this ever-growing fear that AI and robots will take our jobs and unleash mass unemployment. I think it’s finally time to embrace Ai’s superior prediction power and use it to our advantage.

Tarun Wig, Co-founder, Innefu Labs

Tarun Wig, Co-founder, Innefu Labs

From self-driving cars to virtual assistants, artificial intelligence(AI) is advancing like never before. So far, all we got is narrow AI – artificial intelligence algorithms that are designed to perform a somewhat limited task. The long-term goal is to create a general AI – a complete AI solution capable of outperforming humans at almost every cognitive tasks.

Artificial intelligence making tremendous progress in any kind of forecasting activity. It’s ability to continually monitor totally opposite aspects such as natural disasters, political landscapes, and markets shifts give finance organizations and incredible tool for planning revenue more accurately.

Although we still got a long way until we manage to develop a general AI, we are already seeing the benefits of using artificial intelligence in almost any sector you can think of.

Quake alert

Look at the progress scientists have made on predicting earthquakes by using artificial intelligence. World renowned researchers have spent their prime career years trying out different methods of predicting earthquakes, but to no avail. Luckily, machine-learning algorithms are getting increasingly powerful. Scientists are now able to feed an AI with massive sets of data taken before, during and after an earthquake event. Even better, machine algorithms are now able to look for predictive clues and pick up on different predictive markers.

There’s even an app called ShakeAlert that will emit an extremely annoying noise up to a minute before an earthquake hits. While one minute might not seem like much, it’s enough to be a game-changer for people living in quake-sensible areas. A team of scientists specialized in earthquake prediction states that AIs are well on track in predicting quakes months or even years ahead of time.

AI in Cybersecurity – Friend or Foe?

Artificial intelligence already exceeds the smartest human’s capabilities by leaps and bounds. Imagine a human being filtering through massive amounts of information while keeping on top of all that data – It’s simply not possible. On the other hand, an AI can handle all that information without too much effort on a 24/7/365 basis.

But some experts believe the rise of artificial intelligence will lead to an increase in social engineering attacks, network penetrations, and data theft. Believe it or not, our best defense against Ai-enabled hacking is by using security AIs. It all depends on which side they are on. An AI can be used to defend a cyber-infrastructure, but it’s just as effective in increasing the attack surface that hackers can target.

Governments, private sector businesses and a lot of big corporations are deploying AI systems to help with cybersecurity defense. Hackers are bound to display some abnormalities when trying to hack into a system. A defense AI is capable of scanning for those behavioral abnormalities that a human being is almost sure to miss.

Unfortunately, every system can be exploited, AI’s included. Cybersecurity is a constant battle where hackers will always hammer at every system. But as hackers adjust to AI systems, cybersecurity experts deploy new countermeasures meant to keep them at bay. It’s a constant cat and mouse game where AIs are seen as a huge asset in the war for digital data.

New tricks in supervised learning

Speaking about predicting the future, there’s already an effective deep learning AI system capable of generating the next frames of a story by analyzing just one image. This AI system is basically able to predict how a scene will play out and imagine an intuitive vision of the immediate future. Impressive, isn’t it? To develop this AI, the team behind it feed their artificial intelligence algorithm with an impressive 2 million videos with an approach called adversarial networks. The videos are known to contain anything from train stations to golf courses. Even more, not a single video was named or labeled in an effort to prevent the AI from cheating.

Such a system is, in theory, capable of figuring out when someone is about to get hurt or even predict a potential mass agitation by picking up on the clues. Some researchers believe the next step in AI development is giving them a basic ability to predict the future.

Emphasis on AI Safety

Looking at the bigger picture, there’s a coordinated effort on keeping AI’s impact on society beneficial. A lot of big names like Stephen Hawking, Steve Wozniak, Bill Gates and Elon Musk have expressed their concerns in the media regarding the risks of building the wrong type of AIs.

Many AI milestones that experts believed were decades ago have been reached in the last 5 years. This made a lot of experts consider the sci-fi possibility of achieving superintelligence in our lifetime. While most experts still believe that human level AIs are a century away, others are betting it would happen before 2060. Setting the future aside, it’s clear that AIs have many years left of helping us before they rebel and take over the world, so it’s best we start making the most out of them now.

Still not convinced of AIs capabilities? Here’s another artificial intelligence algorithm that is predicting heart attacks much more accurate than real doctors. How about this AI that accomplishes 1.800 trading days’ worth of analysis in just 10 minutes? The list is virtually endless, so I think it’s high time we embrace AI’s superior prediction powers.

No Comments so far

Jump into a conversation

No Comments Yet!

You can be the one to start a conversation.