The concept of artificial intelligence (AI) has been on the periphery of mainstream consciousness for decades now. As the tech curve really ramps up, AI is working its way into our newsfeeds and everyday lives more often.
You might have even seen a couple of recognisable technological fellows sharing their opinion on the subject recently.
It sure got my attention.
The Internet: Quick! Two uber-rich-and-famous Silicon Valley guys are arguing like schoolkids on social media….
Me: Ok. Weird, I guess?
The Internet: It’s something to do with artificial intelligence and driverless cars and the safety of the entire human race and a robot apocalypse!
Me: Robot apocalypse, you say?
The Internet: and get this – Ol’ Musky reckons The Zuck has ‘a limited understanding of the topic’!
Me: Alright, alright, I’m up.
This blog is going to tackle 4 concepts relating to artificial intelligence:
- #1. What the heck it is and what the heck it is not
- #2. What to expect from AI in the future
- #3. Why it’s going to matter to you (and your business)
- #4. How you can prepare to make the most of it
So, here we are. Artificial intelligence is causing Twitter beef between billionaires. We have reached peak 2017.
This tasty little back and forth between Facebook founder Mark Zuckerberg and Tesla/Space X head honcho Elon Musk was enough to raise more than just the furrowed brows of the tech community. Are we headed toward a “Singularity” where computers surpass all human knowledge and intelligence? We’ll come back to that.
For now, it’s safe to say both Musk and Zuckerberg each have substantial commercial interest in bringing artificial intelligence to the masses, but what on earth are they actually on about?
For the world’s biggest social media network, artificial intelligence is an integral component in the future growth of the platform. Check out the 2017 Facebook F8 Developers conference for a sneak peek at what’s in store.
In camp Tesla, confidence continues to grow around the autonomy of the driverless car.
What’s strange about this little stoush is that our heroes, Zuckerberg and Musk, appear to be diametrically opposed on the issue of artificial intelligence.
On the one hand, we have the warm and fuzzy utopia of Zuck’s AI world, which enhances our social experiences and makes us even ‘more connected’. Definitely no supercomputers enslaving the human race to see here.
Flip the channel and there’s Musk calling for proactive government regulation to slow the development of AI systems and casually reminding us that he “keeps sounding the alarm bell, but until people see robots going down the street killing people, they don’t know how to react.”
Whilst a robot apocalypse is highly unlikely in our lifetime, there is plenty of evidence to suggest that we are actually pretty close to a true technological revolution. An AI revolution could potentially relieve the stress on our tired little minds in the same way that mechanisation gave our aching backs some reprieve during the Industrial Revolution.
But, when someone as influential as Elon Musk is warning us about the future safety of humanity and the need for government regulation of a technology, a few ears are bound to prick up.
If you are like me, the concept of artificial intelligence is stored in a special part of your brain reserved for Sci-Fi Mumbo Jumbo. With pop-culture reference points like The Terminator’s Skynet, Johnny’s Depp’s character in Transcendence, or the android uprising in I, Robot, it’s easy to let our imaginations get the better of us.
But how much of what we see in films and television is an accurate reflection of the state of artificial intelligence in the real world?
Honestly, not that much.
I think one of the reasons it is hard for people to get excited about AI is because our exposure is typically one of two opposing extremes.
- #1. The Hollywood Version – Ridiculously far-fetched and dystopian (but very entertaining!) The problem is we can’t process what we see into any context that remotely relates to our lives.
- #2. The uninspiring beta-mode AI Robot prototype – A clunky humanoid machine that can remember the name of its ‘Master’ but would probably be out of its depth in a conversation with a parrot.
Artificial Intelligence is a pretty hefty topic in terms of breadth and depth. There are entire academic journals are dedicated to artificial intelligence research. However, given the fact that I am not a computer scientist (and I must assume you, dear reader, are probably not either), I invite you to take an introductory leap into the world of AI with me and start scratching the surface of possibilities.
Artificial intelligence? What the heck it is and what the heck it is not
Let’s use the Oxford Dictionary definition as our jumping off point:
Artificial intelligence noun: The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
Within the scope of AI, we often hear terms like machine learning and deep learning. These are essentially subsets within the field of AI, and refer to a system applying a learning algorithm to constantly improve the outcome, based on past experiences.
Humans have a long history of trying to anthropomorphise computer systems. That is, we assign them human-like traits and characteristics in an attempt to make our relationships with technology feel more comfortable.
This is evident in the language we use when we talk about computers. How many times have you heard the following?
- My smartphone is about to die!
- Hang on, the computer is just figuring it out…
- Facebook knows so much about me
- I just let Netflix/Spotify choose for me
Humanising technology is our subconscious effort at trying to normalise a really weird concept. It’s not because we are actually fooled into thinking a robot is human, but rather, our brains are too lazy to deal with the fact that we are interacting with a machine.
In reality, most computer systems we interact with on a daily basis are no match for the broad spectrum of human intelligence.
The function of a basic computer program is simply the result of a command, and will produce an identical result (within the range of programmed alternatives) time and again. Therefore, the value of a computer system is dependent on our ability to operate the functions and interpret the output.
Humans, on the other hand, are extremely complex organisms. We experience emotion, we rationalise based on context and we draw on memory to inform our behaviour and perception in the present. It is these types of processes that computer scientists seek to replicate in AI to enhance the power of a computer system. By applying the same processes that occur in the neural networks of our brain, the system starts to learn like a human.
Virtual assistants like Apple’s Siri are starting to incorporate contextual responses based on GPS locations, however it is likely that many of the seemingly witty responses we currently receive are part of a big pool of pre-programmed, albeit sassy, replies.
To illustrate the theoretical difference (in practice this would be extremely difficult to achieve), let’s imagine we have built two Robot Chefs:
Robot #1 is a non-AI robot. It can perform certain tasks in the kitchen and has been programmed to bake the perfect cake.
Robot #2 is an AI robot. It’s also pretty handy in the kitchen, plus it’s able to train itself on enormous data sets.
You have a birthday party coming up and the guests will arrive this afternoon. You fire up Robot Chef #1 and Robot Chef #2 and let them know it’s Suzie’s birthday and you need two delicious cakes.
Robot Chef #1 collects the ingredients and gets straight to work on a chocolate mud cake, except someone has put the salt in the container that usually holds the sugar. Robot Chef #1 puts an entire cup of salt into the mixture.
Robot Chef #2 knows that Suzie’s husband Chris is allergic to chocolate because last time he visited, he had to skip dessert. Since it’s a hot day outside, Robot Chef #2 also decides that the guests might like something a little lighter than a mud cake, and opts for a Pavlova with fresh fruit. The sugar looks slightly different to Robot Chef #2, and a quick test reveals that the salt and sugar have been reversed.
Robot Chef #1 cuts up the mud cake, but it’s a salty disaster. All of the guests choose a slice of the Pavlova. Robot Chef #2 approached the task in a similar way a human would and this ability to apply knowledge in context earns it the winners badge in the bake-off challenge.
What should we expect from Artificial Intelligence?
Just like the divide between Musk and Zuckerberg, the world seems simultaneously excited and apprehensive about what the future of AI holds for us.
Eliminating mundane or labour intensive tasks seems like an obvious win for humanity. But if computers can replace humans, will people slowly start to lose their jobs to automation?
Then there is that deep, dark doomsday fear that maybe a super-intelligent computer could go ahead and create an even more powerful version of itself and so on, until we are overpowered entirely. For this situation to come even close to reality, an awful lot of powerful people would have to take some extremely irresponsible actions.
The reality is that at this stage, computers require massive system capacities to even come close to resembling specific fragments of human intelligence. One of the biggest hurdles for the future of AI according to University of California computer scientist Stuart Russell, is the idea of abstraction and hierarchical thinking in thought processing.
Scientists have successfully designed a system that used deep neural networks and reinforced machine learning (named DQN) to play a simple Atari computer game. Within a few hours, the system was able to develop sophisticated strategies superior to the best human players.
But here is the issue we face at the cutting edge of AI technology today. As Russell puts it:
“We don’t think programs like the DQN network are figuring out abstract representations of actions. There are some games where DQN just doesn’t get it, and the games that are difficult are the ones that require thinking many, many steps ahead in the primitive representations of actions — ones where a person would think, “Oh, what I need to do now is unlock the door,” and unlocking the door involves fetching the key, etcetera. If the machine doesn’t have the representation “unlock the door” then it can’t really ever make progress on that task.
AI systems can do some incredible things, but in terms of recreating the full spectrum of human intelligence, we are still a long way off.
We can still expect some game-changing innovations to arise in the near future thanks to artificial intelligence and machine learning.
Some of the key areas we can expect progress are:
- Natural language processing for conversational interfaces and accurate translation
- Autonomous cars
- Mass automation of data analysis and interpretation
Creation of news and media (Take this test to see if you can tell whether the headlines were written by a human or a computer!)
Why does it matter to businesses?
Gaining a competitive edge is at the forefront of every business strategy. It’s likely that we will see artificially intelligent robots replacing some of the human workforce, including areas like manufacturing, finance and accounting, customer service and even medical surgeons!
Where artificial intelligence systems will provide organisations with enormous benefit is by finding meaningful patterns in data. The revolution for nearly every aspect of business, big or small, is having a machine learn to identify the salient aspects of Big Data and act accordingly.
In turn, this can inform aspects such as customer experience, research and development, sales forecasting, threats from competitors and foreign investment.
Analysing large and complex data sets, even with powerful computational programs, usually requires a human pilot to interpret the output and apply it in a context that provides value to the organisations. This is labour intensive and requires a skill set that many of us may not possess. Hence the salaries of analysts, market researchers, economists and data scientists are usually at the higher end of the scale.
Handing this over to an intelligent system that can follow the decision processes of a middle manager would signal the start of a global artificial intelligence arms race, where the strongest and best equipped organisations would flourish.
What you can do to be AI ready and take advantage
Becoming an AI-friendly organisation requires something of a culture shift to occur. Depending on what level of exposure you and your team have had to the first wave of smart machines, the concept of process automation may be something that requires structured training for both staff and management.
Understanding and education are critical to the adoption of new technology, so it pays to keep it front of mind across the the board.
In fact you may already be reaping the benefits of basic AI without even realising it yet. By optimising your Facebook ads or Google AdWords campaigns using automated bids and budgets via a system that can process data to the point where it understands that if bids go up, budget needs to be allocated toward the best performing ads is simple AI in practice. It’s taking data based on performance metrics and using it to enhance the performance of your campaign, whilst you are out to lunch with a client.
Discussing the capabilities of AI in future planning and long term strategy meetings as well as more informal settings is a good way to introduce the topic without intimidating people or sending a shockwave of panic through your organisation. This can grease the wheels of innovation and get people shifting their mindset about how technology can be integrated into everyday operations.
We definitely don’t recommend walking down the hallways announcing the omnipotence of our evil robot overlords just yet.
Whether we like it or not, artificial intelligence is going to play an increasingly important role in our personal and professional lives over the next few years. I hope that you have enjoyed learning a little more about the capabilities of AI, as well as some of it’s future applications and why it’s important to keep in mind as we hurtle into the future.
If the world of artificial intelligence has captured your imagination, go ahead and add these to your reading list for some fresh perspectives: