Artificial intelligence and the future of SEO

Davide July 5, 2016
34
shares

Google engineers are working on a search algorithm so complex that they themselves don’t understand it.

Discover how #Google is working on an algorithm so complex, that they themselves no longer understand it. #SEO
Click To Tweet

There is a small anecdote about Albert Einstein. He took a bus ride one day. After buying a ticket, the bus driver gave him his change. Einstein checked the amount. Once, twice, three times… he was left with a different amount each time he counted, so he began to count it for the fourth time. The bus driver just sat there in wonder. “What’s wrong with you?” he finally asked, “you never learned how to count?”Einstein chuckled. “I did,” he answered, “but it was not quite my best subject.”

 

 

Computers are great at mathematics, but they are far from intelligent, let alone brilliant. Regardless of how awe struck we are about computer performance, we as humans still have reason to feel superior. Computers, after all, cannot think for themselves. You have to spoon-feed them. Even a chimpanzee has more mental capacity than our trusted mechanical devices.

 

Let’s enjoy this feeling while we still can, as artificial intelligence is on the rise. Not too long from now, computers will outsmart humans in every facet of ability. The possibility that Google will be responsible for this breakthrough is something that must be seriously considered.

 

Google wants a machine of such wisdom that it could solve all problems faced by mankind. Problems that include disease, death and driving cars. And yes, even problems when it comes to searching on the Internet.

 

This article will first center on the ways in which Google works on the implementation of this dream.
Then, we move on to the consequences for human kind and the opinions of several experts.
And finally, we speculate about the consequences for that profession in which we still can call ourselves experts 🙂

  1. Deep Learning
  2. Artificial intelligence: intelligent and wise, or intelligent and dangerous?
  3. SEO in the era of the wise computer

Deep Learning

Aside from all the great plans Google has for humanity, the objectives of the company can be laid down as follows:

  1. maximize profits from advertisement revenues
  2. get visitors to click on advertisements
  3. receive as many visitors as possible
  4. by answering queries much better than other search engines

How do you begin to perfectly answer three billion search queries a day? Outsourcing seems to be the best option. To a machine, with the capacity and speed of computers combined with Einstein’s genius. A machine that views the world as humans do, but on top of that a machine that has the superhuman ability to know all there is to know and cough those facts up from memory within fractions of a second. This has been Google’s goal from the very start. Real artificial intelligence, much smarter than what’s currently available. To get a concept of what that intelligence might look like, we first have to define the word intelligence.

 

What does intelligence mean to us?

Intelligence has many definitions. For our discussion here, the next one fits nicely: the capacity to learn independently and respond to new challenges flexibly. If we translate this to artificial intelligence, it involves a computer that is no longer dependent on human input, but rather one that can think and act on its own.

 

Any idea what this sort of computer would look like on the inside? There’s an analogy very close to us: the human brain. In the 1980’s, scientists began to experiment with ‘neural net’ systems, which simulate the human brain’s learning process. One of its pioneers was Geoffrey Hinton, a researcher affiliated with the University of Toronto. He also works part-time for Google. He is one of many artificial intelligence researchers enlisted by the company in the past decade.

 

Deep Mind

A giant leap forward in the development of the wise computer was taken by Google in 2013 with the acquisition of DeepMind, a UK Artificial Intelligence startup, which also garnered the attention of Facebook. DeepMind was founded several years prior by Demis Hassabis, a former chess prodigy who had earned his stripes in the games industry.

 

Before the announcement of the take-over, few had ever heard of DeepMind. The small company kept itself away from any publicity as they worked on a remarkable project, which they described as: the passive training of a neural net to play Atari video games.

 

DeepMind’s neural net system was left to its own device for millions of games played. It had to independently learn the rules and determine a winning strategy. For an expensive computer, that might sound a bit unproductive. It played Pong, Space Invader, Beam Rider and other simple Atari classics at the level of a talented teenager, but the intriguing part was that some of the system’s most successful strategies were entirely novel. No man had ever considered playing like that. As Hassabis put it: “This highlights the truly remarkable potential of deep learning technology. We provide the system with the ability to learn from experience, just like we humans do, and then it might conjure up solutions that we possibly could never even pre-program. It’s exciting to see it play an Atari game with strategies the programmers knew nothing about.”

 

It’s definitely exciting, not as a final result, but certainly in terms of potential. In a bit we will see what new horizons deep learning opens up for search engines. But first, let’s see if we should be that pleased with these developments.

Artificial intelligence: intelligent and wise or intelligent and dangerous?

Picture an intelligent machine. An artificial super brain that knows all there is to know and keeps on accumulating knowledge by the second. Knowledge far beyond our grasp. What would its personality be like? Would it act like a caring mother, doing what’s right for her children? Would it exhibit friendly behavior, with a philosophical outlook on life? Or would it be a tyrant, bent on absolute power and unmoved by principles or empathy?

 

This is just science fiction drama, surely? Fearmongering, at best? A machine is devoid of any personality, and you can always just pull the plug on it. Or can you?

Several leading figures from science and the IT industry have been quick to warn us about the potential risks of artificial intelligence. Here are a few quotes – and not just from anyone:

attlefield robot
Battlefield robot by Dick Thomas Johnson (licensed under CC BY 2.0)
 

Stephen Hawking:

Stephen Hawking
Stephen Hawking in Cambridge” by Doug Wheller (licensed under CC BY 2.0)

Tesla CEO Elon Musk:

Elon Musk
Elon Musk by Dan Taylor (licensed under CC BY 2.0)

Steve Wozniak, co-founder of Apple:
“Like people including Stephen Hawking and Elon Musk have predicted, I agree that the future is scary and very bad for people. If we build these devices to take care of everything for us, eventually they’ll think faster than us and they’ll get rid of the slow humans.” He added:

#AI Will we be the Gods? Will we be the family pets? Or will we be ants that get stepped on? #Wozniak
Click To Tweet
Recently Steve Wozniak has answered his own questions. He is no longer afraid of the future, he said, because “Robots will look upon humans as their family pets and take good care of us.”

 
Steve Wozniak
Steve Wozniak by Al Luckow

Bill Gates:

Within a few decades, #AI will be strong enough to be a concern. #Gates
Click To Tweet
Global Economic Crisis: Role and Challenges of the G20: William H. Gates III
Bill Gates by World Economic Forum (licensed under CC BY 2.0)

 

But of course, there are people who only see the advantages of artificial intelligence.

 

The renowned spiritual master Osho called it a blessing back in the 1980’s and considered all-knowing computers as a necessary condition for the rise of ‘the new man’. People always fear novelties, he said, but what is there to be afraid of? Computers have no religion and do not meddle in politics, they do not wage war and they do not discriminate. Because they will do all the thinking for us, we can spend our time dancing, singing and meditating. Laziness will become a spiritual quality for the first time in history.

 

Does Google heed the warnings? Is there awareness about the risks instead of just a fascination with the possibilities of a gigantic artificial brain that absorbs all knowledge, processes it independently and takes decisions based on that knowledge that are far beyond our understanding?

 

We know that Google imposes certain restrictions on itself. As a condition for the acquisition of DeepMind, Hassabis and his partners set the demand that an external Advisory Board should supervise the progress of Google’s artificial intelligence projects and that the technology must never be sold to the army or intelligence agencies.

 

The fear that artificial intelligence could spiral out of control, and the guidelines that prevent it, are reminiscent of the famous three laws of robotics by Isaac Asimov.

 

And this takes us back to science fiction. It remains nothing but speculation. So what prevents us from speculating about the impact of an all-knowing, curious and hopefully wise artificial brain on search engines and SEO.

SEO in the era of the wise computer

Let’s switch back to what DeepMind’s founder Demis Hassabis had to say. “This highlights the truly remarkable potential of deep learning technology. We provide the system with the ability to learn from experience, just like we humans do, and then it might conjure up solutions that we possibly could never even pre-program.”

 

Here’s the fun part. If we continue this train of thought, we end up with a search engine capable of providing results that the clever folks at Google could have never have programmed. And in the next stage of its learning process, these clever people can kick back their chairs entirely, as the machine will take care of its input by itself.

 

Can you picture it?

Question to Google: Hey, I’ve suddenly reached the top position with my homepage, how did that happen?
Answer: No clue.
Question: How important are links in your algorithm?
Answer: Beats me…

 

The sunny side in this future vision is that Google will bother us a lot less with updates. Even though the brain will update its algorithm every second, no one will be able to grasp the how or what, making an explanation rather difficult.

 

Another pleasing consequence is that websites that succeed in satisfying the visitor will rank even higher. This is already a visible development, but one that will be more noticeable with the advent of the artificial brain.

 

And the question: “How do I open the lid off this can?” will finally be answered satisfactorily.

DIN, Google’s “strange questions” project

People from all over the globe pose some 3 billion questions to Google every day. But that’s just a small percentage of the questions on their mind, which they refrain from asking Google. That might all change when the search engine’s artificial intelligence matures.

 

Google is not entirely clueless about the types of questions that might come their way. In 2011, the company founded a project named Daily Information Needs (DIN). Its purpose is to find out what things people ponder about during the day. The things they genuinely want to know, whether Google has the answer or not.

 

It goes as follows: in line with Google’s principle to get the crowd working to its advantage, the company found a large number of people willing to share their information needs upon demand. At 8 random intervals during the day, these people get sent the following question: what would you like to know right now? Towards the end of the day, they fill in a list with their questions and mention if they have tried to obtain an answer, what they did to get these answers and whether they’ve succeeded.

 

Some questions allow for categorization and help Google to set priorities. Many questions are of the “how to” type and the results are used by Google for several purposes, including ranking these “how to” videos on Youtube.

 

That aside, there are also plenty of questions that Google has no answer to as of yet. Questions like: “When will my husband finally grow up?” “Where can I quickly get 200 dollars?” “At what time does the meeting start?”

Another complicating parameter is the divergent intention with which people from different cultures ask the same question. When an American asks: “When does the train arrive?” he wants to know if there’s been a delay. A Japanese person, however, is used to the trains being right on time, but he wants to know at what time he should leave his house to make his way through the rush of the station.

 

What can an artificial brain do to answer these kinds of questions in the future? We don’t know. But it will offer solutions we have never even considered. Factors that have been given zero thought thus far, strange alien correlations, an unfathomable algorithm that results in the perfect answer to the question: “How do I open the lid off this jar?”

 

A wise search engine will answer queries with a deep understanding of what the person posing the question truly wants. In fact, the search engine will know this person better than he or she knows him/herself. And when it comes to SEO, Consumer Experience might just end up being the only parameter left to work on and fully understand.

 

The wise search engine is coming, and it’s best to anticipate on its arrival ahead of time. But in the meantime, we should also keep our eyes on the road and perform SEO according to 2016 standards. Less speculative and immediately applicable SEO advice can be found in the free ebook How to predict the future of SEO.