When we say AI, many of us don’t actually mean ‘Artificial Intelligence.’
Instead, what we would probably want to say is IA – ‘Intelligence Augmentation.’
So what’s the difference between the two?
I. Artificial Intelligence vs. Intelligence Augmentation
Let’s start by looking at Artificial Intelligence (AI). ‘Artificial Intelligence’ refers to the idea of building human-like machines. In other words, these machines are designed to reproduce human cognition and function autonomously like an actual human being would. They accept inputs and modify their control behaviors as they ‘learn’ gradually. Completing the control loop is the critical component to true AI.
A true example of Artificial Intelligence is the Google Self-Driving Car. It uses sensors and software to navigate through complicated scenarios on the road. Then, it produces a set of possible solutions, picks an option, and acts upon it. The Google Car works alone and solves problems by itself.
Advances such as these show that the field of AI has undeniably made tremendous progress over the past few decades. Five years ago, IBM Watson – a question answering machine – won a game show called ‘Jeopardy’ against humans by its ability to store massive information, process questions presented in natural language, and draw answers. More recently, Google’s Deepmind AlphaGo competed with the world’s top Go-player. If Moore’s Law persists, this will subsequently increase the possibility of machines reaching human-level computing power in the future.
While the advancements of AI are indeed exciting, there’s still discomfort from the public that AI will reach beyond human capabilities, replace us from our jobs, dominate us, and soon make most human economic functions obsolete. That’s not an irrational fear considering the constant exaggeration in the media and fictitious Hollywood productions have made us believe. However, that train of thought is a far cry from reality and should not result in us pulling the plug from making technological advancements.
But why build intelligence from scratch when we can amplify what’s already there?
Some wondered, and another concept emerged: Intelligence Augmentation (IA). It was even referred to as ‘the new AI’ in a paper published by the Institute of Electrical and Electronics Engineers from 2012.
Unlike ‘pure AI’, Intelligence Augmentation is designed to build machines that leverage the human mind and capabilities that have been proven to work. The concept does not intend to replicate a human’s mind, but rather to improve it.
For instance, think about search engines. Search engines allow us to find what we need in a matter of seconds. With algorithms that are constantly being improved, the process only gets faster and more accurate, which allows us to perform research much more efficiently. This is a stark contrast to the olden times when we would take days going over hundred of books in libraries or looking through yellow pages to find information. The search engines now do the grunt-work (collecting data) for us, speed up the process, and allow us to move on to more advanced tasks (analysis and discovery).
In short, the distinction between Artificial Intelligence and Intelligence Augmentation is that in AI, the machine acts upon its own decision, whereas in IA, the human aspect is still the heart of the process.
II. Humans + Machines = Unparalleled Combination
Historically, the symbiosis between humans and machines has proven to still be the most powerful combination of all.
In 2005, Playchess.com hosted a tournament in which anybody could compete in teams with other computers or human players. The winning team comprised of two amateur chess players and three computers, beating their opponents who were either grandmasters or strong computers, or a combination of both.
Let’s take a deeper look into the results, as described by Garry Kasparov in 2010:
The Grandmasters with a relatively weak PC were able to beat Hydra, a chess-playing specific machine similar to IBM’s Deep Blue. With that result, human capabilities augmented by machine intelligence still proved to be the most powerful.
However, the two amateurs with three computers of average computing power were the last ones standing. They didn’t have the skills of the Grandmasters, and neither of their machines possessed significant computing power. How did they end up winning the entire game then?
“Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process.” (Garry Kasparov, 2010)
The two players were able to coach the three computers to look for insights and make the best moves. They knew how to leverage technology to its best use. In the end, it was not about the most skilled human or the strongest computer, it was about knowing how to best solve the problem with the provided resources.
In the field of technology development, we call this making ‘algorithmic advances’. The software with the best algorithm usually turns out to be the most powerful.
III. What’s becoming of these technologies?
Although it might, or might not, come as a surprise, we have effectively been using IA in our everyday lives without realizing it.
Search engines, as mentioned above, are one example. Do you know that your Facebook or Instagram feeds are also a form of Intelligence Augmentation?
Facebook’s newsfeed, or any sort of content curation provider, goes through all available content to present, makes selections based on your set preferences, and shows you updates from the people or the types of content you care about. The algorithm filters an ever-flooding stream of updates for you, so that you can spend time looking at what is truly valuable to you.
Siri, Cortana, Google Talk, or any virtual assistants out there help us speed up the mundane daily tasks and let us spend our time much more productively.
A majority of technologies that we are working on today would more or less be classified as ‘Intelligence Augmentation’, which are designed with the intention of furthering human capabilities rather than replacing it. With these advancing developments, we are allowed to move onto higher order tasks, achieve bigger goals, and solve harder problems – just like we have always done.
