For the last couple of years AI technology has remained as one of the most discussed points. And there is no surprise in that. While some, as Microsoft's chief envisioning officer Dave Coplin called it "the most important technology that anybody on the planet is working on today", others like Stephen Hawking and Elon Musk are concerned that AI can pose a threat to humanity, if it is developed in the wrong way. And since AI is not only one of the most rapidly developing technologies up to date but also raises such opposing opinions, let's try to understand how Artificial Intelligence can shape our world in the future.
As good as our brain
Firstly, we need to establish a clear understanding of AI, and how it operates. In a nutshell, AI should mimic the work of a human brain. But we, as humans, aren't born intelligent right away - we become intelligent by observation of the environment and proper reaction to it. Same is true for the AI technology. It cannot mimic human brain straight away without training.
As seen on the image above, AI itself is actually the inner circle in the whole brain mimicking process. AI cannot learn itself, without a Machine Learning (ML), which in its own turn relies on Deep Learning (DL). But for all these three areas to work properly, the deepest and the most important area should function - Neural Networks (NN).
AI only tries to mimic human brain, where Neural Networks (should in theory) function like one. Our brain, spinal cord and peripheral nerves make up a complex, integrated information-processing and control system, that eventually helps us to think, dream and reason. So how close modern Neural Networks are to this level of processing? The answer is - they still have a pretty long way to go.
Today's NN is at its best, when it needs to complete a narrow and straightforward task. We can already see how it's successfully implemented in various aspects of everyday routines. Netflix will advise you the most relevant movie, better than your friend would, and Google Maps will find the best route to your destination without traffic. This is why Convolutional Neural Networks (CNN) was one of the key trends in 2017, with breakthrough abilities in computer vision and natural language processing. But with AI technology developing incredibly fast, CNN has already proven to be very limited in its abilities. Just remember the hilarious MIT researchers experiment, where they confused Google's object recognition AI into thinking that turtle is actually a rifle.
We hope that recently announced Capsule networks (CapsNet) will help to fix this problem, and push AI based program abilities even further.
Knowing all that we can safely say that the more AI will resemble human brain, the more knowledge we will receive about the last, pushing limits of future AI even further. This type of interconnected relation is something that will inevitably lead humanity to "that type of future" we all see in fiction movies and novels. And although discussions of cybernetic robots and flying cars sound very intriguing, let's focus on more realistic, nonetheless exciting, close future.
Future is closer than you think
Improved social services will be something you'll notice first. They will become smarter, more automatic and responsive. How about Instagram automatically tagging all your friends on photos, saving you from doing this routine by yourself, sounds pretty cool, right?
Personal assistants, like Siri and Alexa, will recognize and understand your voice better. Last is known, for already being trained to hear suicidal notes in your speech. This means that future assistants will not only hear "what" you say, but also "how" you say it.
Regarding new safety options: imagine a self-driven car that will automatically stop at any speed if too close to an obstacle. Modern scanning and image reading programs can be applied in a variety of ways, and this one can save many human lives.
Also, healthcare will definitely become a huge winner, being able to give more early and accurate predictions on symptoms. Some companies are widely using the technology in the healthcare software development. Risk of getting ill can be also covered in different areas of work - it will be more likely to spot substandard or dated food product on every step of production.
Entertainment will also transform into something completely brand new, bringing types of content that were unimaginable before. How about a competition between two AI programs in any type of game or activity? And God, who knows what those Dating shows will try to experiment with in the future.
Let's also look at some interesting AI related news that happened during the last year:
Researchers at MIT presented their AI cybersecurity platform that could detect cyber attacks with accuracy of 85 percent.
Japanese advertising agency will employ "world's first" AI creative director to help them craft TV commercials.
Social humanoid "Sophia" became a Saudi Arabian citizen, becoming a first ever robot to ever get a citizenship.
And some AI in numbers:
- AI will replace 16% of our jobs over the next decade
- 80% of the firms will rely on "insights-as-a-service" in 2018 at least in some capacity
- AI will manage 85% of customer interactions by 2020
- Projected size of AI market will be $36.8 billion by 2025