Mark Zuckerberg's work is never over. Fresh from his two-month paternal leave, the Facebook founder and CEO is already proposing on building a state-of-the-art artificial intelligence.

Earlier this month, Zuckerberg revealed on his New Year's resolution that he wants an AI technology that will help him with his work and serve as his home's "virtual majordomo."

A couple of days ago, he updated his plans and discussed the current limitations of artificial intelligence that he thinks need to be addressed soon, CNBC has learned.

According to Zuckerberg, "Simply put, today's AI is good at recognizing patterns and bad at what we would call common sense."

Currently, artificial intelligence uses pattern recognition to "learn" things. For example, if someone wants to train their AI to recognize their face, he must show it his picture repeatedly until it recognizes a pattern.

But for Zuckerberg, this technology has great limitations. He cited that when a real person learns something new, they do not need to see it multiple times in order for them to recognize a pattern. For him, the current technology employed by artificial intelligence is still lagging compared to actual humans.

Zuckerberg suggested that for an AI system anticipate and react to new situations quickly, he said that a method of unsupervised learning must be used. This type of learning is how humans and animals observe and learn from their surroundings, leading to a better and more appropriate reaction.

In a related report by CNN, Zuckerberg noted that some people thinks that this version of the AI are just a few steps away. He argued that is not the case as experts have yet to utilize or at the least, understand how general learning works.

Another hurdle for Zuckerberg's vision are the huge number of tech icons that are not too keen on a very smart AI system.

Among those who see AI as a threat are Microsoft founder Bill Gates, Tesla founder Elon Musk and famous physicist Stephen Hawking.

But Zuckerberg countered that people should not be afraid of artificial intelligence, but instead hope that it will be developed soon because it has the potential to do a huge amount of good in the world.

For him, once a faster, smarter and better AI system is invented, it can help save lives by diagnosing diseases and even be used on self-driving cars.

He also countered that it can also help with humanity's effort in space exploration because it can reach places mankind cannot.

As of now, humanity is banking on J.A.R.V.I.S. instead of SkyNet.