One Ring to rule them all, One Ring to find them,
One Ring to bring them all and in the darkness bind them
Was the above quote about AI?
Seems like Elon Musk has taken a fatalistic attitude towards AI. While talking to Joe Rogan, he mentioned speaking to Congress and many tech leaders — urging them to uphold or at least regulate AI development. As per the founder to Tesla and SpaceX, no one listened, and the world might be heading towards a future where machines will be in charge.
Elon further added that human intelligence won’t be able to compete with the future machines and all we humans are doing, via Google and FB, is feeding a “system” with all possible human interactions and tendencies. This system once trained enough might spit out a future, which might be very different from the current one that we are projecting.
But, let’s look at the other side of the argument — the one that proposes a better human existence with AI.
a16z, a leading VC from Silicon Valley, is a huge proponent of AI and is investing a lot in this domain. To them, AI/ML development has arrived and is here to stay and the investment in this domain is on rocket fuel. Their many portfolio companies also indicate their interest, so they are putting money where their mouth is.
Peter Thiel, the founder of PayPal, is of the opinion that most of the people have watched terminator more times than intended. And this has resulted in people hallucinating over a future where robots with AI will rule the world.
If you look at governments, you will find clear direction and narrative towards excelling in AI. China is super focused and public about AI ambition and wants to be the leader in AI by 2030. Putin from Russia also believes that the next ruling nation will rule via competence in AI.
All the above sounds scary and puzzling and does warrant an investigation into what AI is capable of and where it’s heading.
So let’s look into the world of AI. Where is the development is really at? Should you be worried about it? Or is there is a better future coming for all of us.
Narrow vs General vs Super AI
AI can be divided into three stages. Narrow, General and Super AI.
With Narrow AI computers assist people with specific dedicated tasks. With General AI, computers will have the ability to learn from one domain and extend it to another one. With Super AI, machines will be so good that humans will not be able to match their intelligence.
Consider it in a way, Narrow AI is like an assistant to humans, and we use it to make our life easier. With General AI, you are looking at AI and computers as your peers. But with Super AI, humans will become a tiny organism against a much superior intelligent counterpart.
Right now, all the progress that’s been made is in the narrow AI domain. You might have used some of the examples of narrow AI working in applications like Pinterest, Airbnb, Amazon Alexa, language translation (Google, Facebook) and so on. So, when Pinterest identifies a glass in the picture and suggests you similar options, that’s narrow AI at work. When Airbnb suggests you the optimal price for best pricing your space, that’s AI in work.
Many AI startups are working with the narrow AI and developing technological solutions that can self-drive or self-fly or self-identify objects, issues, voice, images and so on.
Is this promising? Yes, seems like it. As with this handy application similar to what databases were, narrow AI is becoming essentiality for all the companies. As per a16z, people will start labeling their products with AI inside soon, similar to what Intel did. Also, to experts, right now AI seems to be an extension of intelligent databases.
AI vs Human Intelligence
All the applications of AI so far have been in the domain of Narrow AI. Now the questions arise, does this compare to human intelligence at the moment?
For more general intelligence AI — mimicking that of a human — you will find AI lacking behind even in the simple tasks common to humans and scientists are worrying that it’s not progressing well. Many experts have come up with their own version of tests, if completed by machines independently, will indicate that General AI has arrived. Some of the proposed tests include making a coffee and assembling furniture and so on. But the current state of AI is far from this stage.
As per a16z, Geoffrey Hinton, widely considered the authority in AI and deep learning, has come to the conclusion that the progress in AI has come to a halt. The backpropagation algorithm that has been attributed to most of the narrow AI advancement can’t take this field of study further. As per him, scientists need to start over.
This is the hottest topic and many experts are worried about it. But proponents of AI compare this to the period when machines took over the farms or robots took over the industrial jobs. But such shifts eventually created more industries and better jobs.
Also, they cite examples where employment increased parallel to an increase in automation. Amazon deploys robots, yet they also hire more employees to work with them.
For now, it seems like humanity and machines will be working together to build a better future. Enough innovation in the field is there to exploit its potential for many years to come. The venue is open for governments, corporations, and entrepreneurs to build products with “AI Inside”.