Where we’re at: Artificial intelligence vs. machine learning

Humanity still has the edge… for now


Despite all the reports of investments and advancements put towards Artificial Intelligence(A.I.) it is still often hard to understand what A.I. is.  For many, A.I. comes in the Hollywood-esque form of Ava, from Ex- Machina, a robot that could easily be mistaken for a human. From the way she looks to her desires, wants, andskills, she is as close to human as it gets. For others, A.I. looks like Apples “intelligent assistant,”  Siri  who can answer questions and operate Google for us.

With advancements like robots connecting to our brains for feedback the possibilities for how we interact with computers are wide open.

Consciousness is a complex system that has become a buzzword among millennials. People are looking to obtain it, create it or understand it. The definition of consciousness is “the awareness by the mind of itself and the world.” It is the idea of artificial consciousness that makes A.I. such a hot and somewhat scary topic, which creates ambiguity. Recently, the CBC reported that the Federal Government will be making an investment of 125 million dollars to the Vector Institute for Artificial Intelligence. Provincially, Ontario will be adding 50 million dollars. That’s 175 million dollars being invested into something that a lot of people are unclear about.

Ravi Chander, has been a software architect for eight years. according to Chander, his job consists of understanding the problems an organization is facing, and come up with solutions to solve those problems.

He clarifies A.I as two different things. “In my view, when people use the word Artificial Intelligence there is a very fantasy-oriented definition of Artificial Intelligence, which is more philosophical than it is technical,” says Chandler. “Then there is another category that is lumped into A.I, which is machine learning and that’s a much more pragmatic approach.”

Chander clarifies that we are in the age of machine learning, which he describes as the ability for machines to sift through massive amounts of information and identify patterns to learn and implement different variations of that information. Whereas Artificial Intelligence, which Chander says can be used as an umbrella term, is more specifically the idea of machines gaining conscious awareness.
It’s fair to say 2017 isn’t the year of machines gaining sentience. So, what does artificial intelligence look like in today’s technological landscape?

A household name in the machine learning industry is Siri. This apple addition can schedule reminders for you, look up the definition of words; find the nearest Tim Hortons, and much more. In a recent article published by the Toronto Star, we see one of the limitations of machine learning. The article talks about a glitch when users asked Siri to find them escorts in Toronto instead sent them to a local Esports bar. With only one letter separating the two words, it is an easy mistake. Machine learning is literally in the hands of millions of people, but it is also making its way into our lives in more discrete ways.

Jasper is a Toronto-based recruiting company that uses machine learning to replace human recruiters with machines. CEO and founder, Omid Aminfar, says they are able to reduce the cost for businesses, and eliminate friction through replicating two main things humans do: make decisions and communicate using natural language..

“A lot of people think that -A.I.- is something that can be added to any product or some people think it is a flavour shot that you can add to your coffee today and now your business is A.I powered. That is not the case,” says Aminfar.   Another concern Aminfar addresses is the idea of robots taking jobs on a large scale which he claims is  a miscalculation regarding the abilities of A.I.

It’s through patterns that the Toronto-based company, Logojoy is able to replace a graphic designer in order to complete simple logos for small businesses, people and brands. CEO, Dawson Whitfield, describes the different components of each logo as ingredients, “ It (the A.I) learns over time and we track everything that our users do. For example, let’s say 100 people change their font to a bolder font—a font with a heavier weight to it,we can track that,” Whitfield says. “So, the algorithm basically learned, okay in the future if they have a very bright colour don’t couple that with a thin font.”

Through recognizing patterns, Logojoy allows users to customize their font, colour, and icons. The best part is the finished product is delivered immediately, something that would be impossible working with Logojoy’s human counterpart. “As a consequence of Logojoy and these other A.I tools, small businesses will actually be able to do design by themselves. Which is really exciting because we are completely empowering people to start businesses and get one step closer to the dream,”says Whitfield.

But like Aminfar, Whitfield is confident there will always be work only humans can accomplishso the graphic design field is not obsolete.

The tone overall is that A.I. is a long way from leaving us in some sort of Matrix,but recent developments out of the Massachusetts Institute of Technology (MIT) tell us the field is ever evolving. In March, MIT published a study about a robot who, through connecting to brain waves, is able to learn and collect data. “What is very interesting about this particular application of machine learning,” says Chander who also has a masters in Computer Science, “is that instead of getting information from external sources, the source of information for these machines is actually human being direct electrical signals.”

According to Chander,up until now it has taken, hundreds of thousands of hours for developers to teach machines the amount of information necessary to create the appropriate knowledge base. After this development, Chander is excited about the real-time nature of the feedback loop. Chander has some ideas that are a little more worldly than logo design and job recruiting when it comes to the future of machine learning.

“Imagine a scenario where you can take something and enhance it with the ability for the machines to learn from the mistakes and now you’ve got a situation where you can send machines into incredibly dangerous situations,” says Chander. “Whether it’s war zones, or whether it’s a post war zone disaster areas, whether it’s nuclear power plants and be able to train them to understand how to look for humans or how to look for enemy combatants. So, the humans are still in the command centre directly linked to the robot but now the human is out of danger.”

Leave a Comment

Your email address will not be published. Required fields are marked *