In the abstract, we all know that artificial intelligence will have a significant impact on our work and business, but in practice most of us just going to be spectators or even prey. How could we all be actors? A few months ago I was in China discussing AI in a series of executive workshops. I was totally focused on the complex managerial discussion surrounding AI. A young interpreter facilitated the English Chinese discussion. After a meeting she
Came up to me and said: could I ask you a question, sir sure I said well, what will happen to me? Amazing, the talented young woman not only translated the discussion but followed the content and started to think about the potentially dire implications for own career. But then, without waiting for an answer she said, and can you tell me how I can
Learn about more about AI and apply it myself bingo. I think we should all be asking ourselves this question far too many people think of AI. As this abstract, some was mysterious force in the hands of experts and stall in set AI is there for all of us to use. In fact, our extensive research with MIT SMR has shown the DES bottleneck for widespread application of AI is not technical expertise. It is a diverse set of people developing an intuitive understanding of AI and putting it to work now. The good news is
A is actually rather simple at its core. However, it has some unusual properties that require getting used to so let us take them one by one. First, you should understand. Ai is an intuition machine. Now we all think of our computers as machines. You probably never looked at your laptop and shout it go great intuition. Now, with the eye, you might one day do just that, because if you think about it, you could
define intuition as acting based on experience and potential you’re having a hard time explaining yourself well that’s precisely what AAA does it learns from experience now how does it do it well in friend and I system the experiences come in the form of fresh beta and some kind of feedback how hes doing so far when learning AI acts like a hiker trying to find the valley in the fog now if you cant see much your best test is to take one little step in the.
steepest direction downwards in the AI world that means addressing your parameters a little bit in the steepest direction towards your goal now you might call it deep learning or back propagation but the basic principles of AI are actually rather simple they then get enhanced with some great ideas and lots of data and processing power now if you take pictures of animals and applaud cats AI learns recognized cats now replace cats by credit card fault or cancer cells and you can see how from simple mechanisms you can start.
Addressing rather complex problems now is this effective because its actually very fast, when machines learned intuition by AI, they didnt forget how to do calculations and its a combination of the two that made Watson win rapidly and deep stack when online poker? You might all know the famous book by Daniel Kahneman, the Nobel Prize laureate, Thinking Fast and Slow how new humans are great in nutrition but, frankly, lousy at calculations. Now, AI enabled machines are good at both
calculations and intuition so might you might describe them as thinking fast and fast now with these properties and lots of determination AI has largely cracked vision and language now this is super important for business because his vision machines can act in the real world giving us robots and self-driving cars and this language they can start interacting with human and accessing human knowledge.
however AI is not a magic wand what you should also know you cannot simply by intelligence you can in fact download almost all naked AI algorithms even for free but they are not natively intelligent that’s they have no business value instead you need to nurture the intelligence by training them on data lots of data often your data a little bit like providing first experiences to.
a newborn child AI is not ready-made the nurturing is required to build the intelligent tool itself now you also should not copy humans why well in my youth I used to play chess I was actually quite good at us I became German youth champion and played Garry Kasparov in the World Championship now you guess the outcome irrespective I was in real pain by I saw Gary become the first world champion interest to lose a match against the machine now this was the big deal for AI because for.
more than 30 years eh I thought if you solve tests you solve intelligence at chess playing ape is smart now when they finally did it the AI community was disappointed they used too much brute force didnt have any inside of beauty and most importantly it proved completely irrelevant to solve anything other than chess but the realist is submarines dont swim what I mean by that is.
machines perform tasks differently from humans a self-driving car should not copy human drivers any more than todays car should copy horses now lets see what we have at base this is simple fast and fast intuition machine with improving vision and language skills whose intelligence you cannot buy but have to nurture with lots of data and feedback and to solve problems differently from humans in other words drop your scare AI is not the complex it can help you find.
patterns even in our real world populated by humans you just have to Train it with lots of data and tailors the approach to AI not people we know from hundreds of hands-on projects and our in-depth research was MIT that these are the core ingredients of applying AI how well you master it will separate winners and losers now I want to take you out of your seats and into the real world some years ago we worked with the head of operations of a Korean copper.
smelter yes a copper smelter it just promised you the real world the guy was having problems for the purity of his copper now his engineers had exhausted their algorithms so he needed to try something new he did have several years of operational data so we joined he sat down and trained his very own intuition machine when we were done the algorithm proclaimed based on my experience I.
would propose to harden the process this way well this modern AI you can get that with or without a dramatic Frankenstein accent now the guy followed the advice the beauty of the copper increased and the company doubled its result now you might think hmm this smelter people were probably really behind the uk-based deepmind has done it to Google now the Google the mother company running the worlds largest AI system in fact it built an algorithm for power.
management that decrease the energy costs at Google Data Center by 15 percent at the Google scale you’re talking real money now tailoring real approach to AI actually does require a specific structure you want to keep actually central but centralize all learning that’s totally different from humans where action and learning lives in the same body and its allows for an unprecedented pace of developing intelligence and scale of applying it heres how it works.
suppose you’re the proud owner of entire network of coffee shops and you want to offer a customer lets call her gene a very special pastry to go with her morning coffee now do not just offer her the average popular choice with AI you can now gather all data from all customers to tailor the very specific offering to the context of gene have you seen the genes of this world love it remember with AI the action happens in the field but the learning happens at headquarters self-driving cars run autonomously but learn centrally now when people and by that I mean all of us.
start applying the knowledge of AI we have seen the applications explode when you are in procurement you can get help reviewing contracts when you run critical equipment you preempt some requirements for a maintenance when you are a lets say recruiter for new jobs you delegate interviewing and scheduling and if your bang happens to be online for cybersecurity you definitely want to recruit a fast and fast assistant now we.
could go on like this forever let me return to our interpreter and allow me a final remark people sometimes worry they might be out of work if they start applying AI but tasks and no jobs and jobs is not work the interpreter was right to think that some of her tasks will be automated but you can redefine her job augmenting her performance with AI and she might even expand her career applying AI to all those Ann tapped opportunities in no.
trans language applications all over the world a AI is not this destructive mysterious force now it was tough but we just went through the procured entities with of AI jointly and with this intuitive understanding we can start to moving from mere spectators or prey to becoming actors in the AI world we become more effective and we can jointly explore new frontiers thank you.
Leave a Reply