The state of AI: 2019
The state of AI: 2019
Artificial Intelligence is one of the fastest growing science at the moment. Although it still existed in the 80s, right now we are living in the era when sci-fi literature is becoming less and less fiction.
From predicting behaviour, analysing sentiment (discerning between happy and unhappy people) and generating realistic pictures of people that don’t exist (see here: https://thispersondoesnotexist.com/) — there are numerous projects (a lot of them being free for public use) that can leave you speechless on the power that computers have in the digital era.
Where are we now?
The machine learning agents called AlphaZero and AlphaGo have gained a lot of attention when they have beaten real champions at chess and at go. These are considered very hard games that require a lot of experience for people to become good at. A human requires a lot of time to attain the rank of an advanced player, never mind a global champion. But in 4 hours or less, machines have risen to a skill level that they used to beat renouned players.
“We have always assumed that chess required too much empirical knowledge for a machine to play so well from scratch, with no human knowledge added at all.” — Garry Kasparov
Go is considered to be a game even more complicated than chess, because of how it becomes more and more complicated throughout the match. The number of possible moves increases with every move made, and it is considered to make use of a kind of developed intuition which only a human can possess with years upon years of practice. But in less than a day, a machine has “learned” enough to become a challenging opponent.
The stories about AI victories in chess and go are not new (they happened in 2016). There are many other interesting simulations where AI has been powering through, using artificially built skill to become a good performer.
One of my favoutite simulations is StarCraft II. Deepmind, the company behind AlphaGo and AlphaZero (the machine learning agents that defeated the world champions at go and chess) have partnered up with Blizzard to create a Machine Learning framework for the well-known strategy game.
Where are we heading?
Apart from playing games, Machine Learning agents are being developed for more hands-on real-world applications as well.
Businesses nowadays are looking for chances of having AI-based tools (that don’t break their whole budget) incorporated in their operations. There is AI-based software that detects fraudulent financial transactions, to other anomalies in heavily conglomerated data that would be hard to follow by mere humans. Software that optimises costs, that chooses the fastest routes from a location to a destination, software that detects diseases much faster and more efficient than humans could.
The trend of AI is permeating all other sectors, not just IT. In a way, it is kind of like the times when computers were being adopted in processes and business operations. Now computers are everywhere. It might be the same case with AI (like movies and literature have been predicting decades or longer). But the key difference is that computers evolved by human research and development, which had to undergo a lot of prototyping, testing, trial and error, and of course budget allocation. AI agents can improve by themselves.
2 years?
It is hard to make any type of prediction with a craft that is growing at such a huge rate. One thing that is for sure, is that this impressive tool, which is Artificial Intelligence has one of its primary purposes to reduce labour, allowing people to focus on their goals and stimulating their creativity instead of hindering it because of physical or time constraints.
NVidia has created an AI agent that created a whole virtual video game world all by itself. With some tweaks in different parameters, humans could shape this world to be in any way they imagine. Simply the fact that a computer program can create a whole world is astounding.
AI is being integrated more and more into medicine. IBM Watson is one example of a framework that is being put to use to advance the medical field and to provide people with speed and efficiency at unprecedented rates. It is quite shocking how big of a technology there is in different parts of the world. In some places machines are taking over human labour to increase accuracy to maximum possible while other places are struggling with electricity, food supplies and basic hygiene.
I cannot give the answer of where AI will take us in 2 years’ time. But technological advancements are hinting that human society is not the same it was 10 years ago, and the coming changes and implementations are most probably going to take us aback even more.
How about 5 years?
Since most cars nowadays have a lot of computers inside them, not even to mention the emerging market of electric cars, and with the internet becoming more and more ubiquitous, automatic driving will surely become a platform for car manufacturers and for road developers to collaborate on. There are so many possibilities here, and so many tools that it is hard to make any kind of prediction.
One thing that AI agents are really good at is predicting and imitating human behaviour, therefore AI-based tools will give marketers the edge when it comes to advertising and getting clients. We will have systems that will communicate like a human, that will make decisions like a human, and will probably be even made to look like a human. We know for sure that applies for the digital world (see link above with web site that generates realistic faces). It seems like Isaac Asimov’s “I, Robot” might not be too far off, but will the future be prosperous and peaceful?
One significant driver of AI development will always be profit. Companies have the purpose of growing and improving in their revenues, which means they have to make more money. How their ethics, their mission and their means to make use of this technological breakthrough which is Artificial Intelligence will affect their actions and the populations around them, only time can unravel. But AI Ethics is a topic for another post,
Conclusion
Sci-Fi writer William Gibson said that “The future is already here — It’s just not very evenly distributed “. The fact that we have tools like never before in the history of man to build extraordinary things that would impact lives by the millions is an understatement. I am very curious to find out what current developments are shaping our society and how this new society will adapt to its new superpowers provided by Artificial Intelligence.
It is said that whatever man creates with AI it will be a reflection of himself and his own intentions. Probably, this gives us, humans, the responsibility of growing and becoming wise enough to harness this opportunity for progress and unity and not for frail and short-sighted goals. Whatever happens, it will impact the whole planet, so we are ALL in this. Together.
