Artificial Intelligence has made incredible progress in the last few years, it almost seems like we're hearing about new AI innovations and use cases every other week.
As more and more technology includes AI, it becomes clear that we are approaching a new era: one where Artificial Intelligence will compete with human intelligence.
Entrepreneur and "real-life ironman" Elon Musk has always been very clear about the enormous potential of AI, and that it could even become a threat to humanity if it's not controlled in the right way.
That's why he has founded the nonprofit startup "OpenAI", that aims to turn AI into something positive and prevent its dangers.
This week, an AI bot from OpenAI managed to show off some incredible technological progress and gained a lot of media attention:
The Artificial Intelligence Bot managed to beat one of the world's best players in the highly competitive battle game Dota 2 in a 1-on-1 match.
The Tournament
The bot by OpenAI demonstrated its skills at the annual live tournament "The International" this past weekend.
It was up against a crowd favorite, the pro player "Dendi" who is one of the best in the world when it comes to Dota 2.
He was defeated by the bot twice, before forfeiting any more future matches.
After the game, Dendi stated that the bot “feels a little like [a] human, but a little like something else.” - many of its moves looked familiar and human.
He was surprised that it was possible for the AI machine to beat a human player - and therefore human intelligence.
'Bot is really fun and challenging to play against.
'I am sure it is possible to beat it. But it has no room for even slight mistakes.'
-Pro Player "Dendi"
AI Learning process
We've seen AI beat humans in games like Chess or Go before - but nothing as advanced as Dota 2.
So how do you teach a computer to play such a complex game like Dota 2?
The key is - you don't actually teach it.
Apparently, the game played against itself over and over for 2 weeks, until it had not only figured out the game, but also the best moves to win over his opponent. Beating a regular Dota 2 bot would only require 1 hour of learning.
The computer managed to learn certain tricks and tactics within the game, and although its actions per minute don't differ much from an average human, it can make the best possible decision within a split second - which makes all the difference.
"Our bot is trained entirely through self play. It starts out completely random with no knowledge of the world, and simply plays against a copy of itself, which means it always has an evenly matched opponent."-Greg Brockman, Co-founder and COO of OpenAI
Risks of Artificial Intelligence
The topic of Artificial Intelligence always also brings some criticism and ethical questions with it.
What if AI turns against us at some point, or overtakes us? How are you supposed to judge mistakes caused an AI, and how are you supposed to hold it accountable for it?
Elon Musk is very aware of this threat and has always voiced his concerns that AI should be regulated "before it's too late".
On friday, after the Dota 2 tournament and his first tweet celebrating the success of the AI bot, he stated on twitter: "If you're not concerned about AI safety, you should be. Vastly more risk than North Korea."
Is Artificial Intelligence a benefit or a threat to humanity?
© Sirwinchester
Two things come to my mind. Awesome & Scary
Its only a matter of time before AI blows up and becomes ubiquitous. Making dreams a reality is so exiting but kind of scary to think about.
In 20 or 25 years human will be robot or robot will lead the world.
If robot can have their own consciousness and some sort of common sense. Maybe.
@sirwinchester, I'm glad you included this quote. ""If you're not concerned about AI safety, you should be. Vastly more risk than North Korea.". With all the alarm bells lately and media pile on, that quote resonated with me a lot.
Same here
I think it's interesting that various people working in AI have very different views on the potential pitfalls.
This article makes a great point: https://arstechnica.com/information-technology/2017/07/zuckerberg-and-musk-are-both-wrong-about-ai/
Musk has also been on about overpopulation recently too, which is a pretty flawed idea. See: https://overpopulationisamyth.com
I also think there is too much emphasis being put on what Elon thinks, if his views on overpopulation are not entirely accurate, what's to say his views on AI will be?
Wow, I didn't think we would be at that stage with AI. I guess the only way to beat is to pull out the USB stick lol.
Are we heading towards a real-life portrayal of i-robot where a guy like Will Smith will be all of humanity's savior? well.. too early to judge I reckon.. Having said that, I don't think beating the best Dota players should be a difficult task for a bot.. soon enough we will have bots for all e-games, not only games, all fields where machine language, NLU and AI is involved to whatever extent. And yes, once the bot becomes super intelligent in extracting the right "intents", "entities" and "actions", they could very well over-ride major logic files that could be a real threat to us.
WOW!! That's incredible!!!!
The future of AI is the future. It will be predominate eventually over everything.
The development of AI is both, exciting and scary at the same time. Only a few weeks ago two bots from Facebook developed an own, more efficient language and now there is a bot who superior over the the human in a complex game. Looks like the I, Robot scenario isn't too far away...
If AI becomes real what is our future?
Your voting will force your voice to give me strength.
thanks.
It is very helpful to us as human beings, because humans have limited physical strength, but unlimited robots. And in this world there is a lot of hard work that can not be done by two human hands.
Some of the thoughts I have seem a little scary. Maybe all those old movies had it right.
I'm pretty sure this is how Skynet came to be! Have we learned nothing from 80s SciFi? @trendhobo
Artificial intelligence will just take us further down the road of dependence on government for survival. Most first world humans have already lost our survival skills. I'd be for going back to hunting and gathering before I would support AI.
Hopefully it creates more jobs in our lifetime instead of elimating them!
Watching this video and seeing the player's reactions was almost scary. The players wanted to keep playing the bot due to the intrigue it inspired in the players, thus further training the bot. It leads me to wonder if we will willingly train AI similarly to the way we train CAPTCHA.
It is just a bot; but even a chess bot can beat a pro player.
Don't be too amazed by "this".... it is coded.
If AI grow quickly, one day they will ruin the world lol
wow this post is very valuable as the information within has great implications.
This post has been ranked within the top 25 most undervalued posts in the second half of Aug 17. We estimate that this post is undervalued by $41.20 as compared to a scenario in which every voter had an equal say.
See the full rankings and details in The Daily Tribune: Aug 17 - Part II. You can also read about some of our methodology, data analysis and technical details in our initial post.
If you are the author and would prefer not to receive these comments, simply reply "Stop" to this comment.
Well done, good point, thanks for this, following you. I hope you have a wonderful day.
So sick thanks for that one !
While this still represents a big point in the development of AI, it's important to realise that the bot had full access to the dota 2 API, so it had access to a lot of information that human players don't. In a lot of cases in practical applications, the AI doesn't have access to anymore info than the human. So the results may not be as good in those cases
The problem is how to control AI run by the bad guy
They will help Humans. And then they will replace humans. And then they will take over our Planet.
-AnarchyEvo
Computational creativity will be the key, not generic AI.
Did not see this coming so soon after it beating the game of Go. Fascinating stuff indeed!
Although Dota 2 is a much more complex MOBA I'm personally not impressed. AI beating Innovation in Star Craft II would be more impressive to me.
It's kind of scary that several of our world tech leaders are so worried about AI. I have seen Elon at several talks recently warning of the potential dangers we are setting ourselves up for. I hope we can figure out a way to keep control of our creations!
Elon Musk is one of the few people prepared to consider the adverse effects of AI.
Out of control AI is not a pleasant thought but that doesn't mean it shouldn't be thought about! We must understand and engage in order to prevent!
This is not scary. We humans loves creating things that can kill us. Just like Guns. Nothing scary.