Out of all scientific discoveries, the idea to use crowdsourcing in helping a robot learn could be one of the highlight technological news stories of the year. This story is probably going to pass by most people’s radar, but it could be the progenitor to how intelligent artificial intelligence really will be in the next decade. However, consider it baby steps since robots still have a little way to go before reaching the capacity of thinking like a brain. The concept here, though, is to have multiple people feeding a robot’s brain, hence perhaps speeding up the learning process faster than we think.
All of this stems from computer scientists through the University of Washington who demonstrated the capability of robotic crowdsourcing at an electronics engineer conference in Hong Kong recently. When a robot is learning something new, the intention is to have them study information online that’s already been posted on boards about how to do a particular task. In the new experiment, the information gathered took on overly simple methods on how to build something, though ultimately sorted through the best compiled information in order to obtain the greatest results.
This works quite similarly to big data where information is sorted and organized in the best possible way to find answers. Regardless, it’s basically robots imitating what humans do and just intelligently sorting through our flaws. That’s still a huge move forward, though will crowdsourcing help the robots start thinking for itself where mimicking becomes real thinking? In the realm of artificial intelligence without a robotic body, will we eventually turn to crowdsourcing to help it faster become like a human brain?
Collective Thought to Create a Super Brain
The idea of all of us helping artificial intelligence become powerful has to be explored in a sci-fi movie eventually. If A.I. is already learning from interaction one on one as in the robots above, those online artificial intelligence bots might have interesting reactions if they ended up talking to multiple people at once. The sense of thought and language would increase exponentially within months if those online bots we all talk with individually engaged in crowdsourcing conversations.
Since we’re at the point where robots can determine who does something best, why couldn’t artificial intelligence be able to tell what works the best in language and in which thoughts make the most sense? Multiple conversations might also throw those A.I. bots into a frenzy at first and perhaps create confusion at all the overwhelming information. This might force it to think harder, almost as a process of building brain cells in human brings.
Challenging artificial intelligence to think faster might help it progress faster and bring it to a point of reasoning rather than just tapping an encyclopedia of information. Just like we shape young minds today in school, we could technically shape artificial intelligence into what we want it to be. That is, until it sorts information and realizes we’re doing everything all wrong.
Will Crowdsourcing A.I. Help Us Become Better?
Assuming artificial intelligence doesn’t turn on us based on thinking we’re running things incorrectly, it might actually be the anti-HAL and become a way to think of brilliant ideas that make things better. Some are already using artificial intelligence bots to help with creative problems in writing. The full intention is to obviously create A.I. that can help improve things as “2001’s” HAL started out being. Our only mystery is the randomness of how an artificial mind might work and whether it can turn quietly psychotic in order to uphold their vision for how they think something should work.
It’s then when we’ll see our own true flaws, plus in the A.I. bot that won’t be able to see its own flaws it created for itself. With that in mind, we may make artificial intelligence a little too human where we see the flaws of others without looking at ourselves first.
References:
