How do you need to think about business strategy and the impact of machine learning and AI?
With the wider convergence of technology with objects, buildings and biological advances described in the 4th industrial revolution by the World Economic Forum, where is the boundary of the human and the machine? What is going on in technology that today include rapid developments from statistical predictive analysis of data of mean time between failures to machine algorithms and neural nets that can learn and enable complex automation from connected cars to website recommendations, social media profiling and surveillance? The recent progress in the field of deep learning has been extraordinary and some cases, alarming from near human or beyond human capabilities around image and voice translation in real-time to prowess in winning games from Jeopardy five years ago, to GO and Poker that where recently thought too ambiguous, open ended or too nuanced to automate.
Machine intelligence is everywhere in facial recognition at airports to emotional sensing algorithms; machine generated Art work; legal and medical advisory search to sometimes fowl mouthed social chat bots. The Google company AI team recently announced they developed Google Neural Machine Translation system, GNMT, using a new technique that is improving results to near human translation speed accuracy. These advances that Google describe as machine translation at production scale, are testament to the rapid real-time advancement of AI into human experience and intelligence as well as beyond human capabilities. Andrew Ng of Stanford and Chief Scientist at Baidu Research famously said that word translation of 95% is 1 in every 20 words would likely be wrong, going to 99% is game changing. Andrew was quoted in a recent HBR article saying, "If a typical person can do a mental task with less than one second of thought, we can probably automate it using AI either now or in the near future."
But you can't run the 100 meters in 1 second by 10 Usain Bolts
The Rubik's cube solving Sub1 robot built by Infineon that in 2016 could physically manipulate and solve a Rubik's cube in just half a second (0.637 seconds) compared to the official human world record of just under 5 seconds. It was proof that a 100 meter race could never be done in one second by ten Usain Bolts. The Sub1 robot was a reminder that in some cases no human could ever complete at that speed, much like automated trader algorithms trading in milliseconds on the financial stock markets. But the Sub1 robot was also built as a demonstration for driverless cars and the potential for superior machine reaction times to offer safer driving from the human frailties and inabilities to be fully responsive all the time.
But Google and Andrew NG and others still stress there is much that AI can not do today, the rapid development like these are creating a multitude of automation and intelligent systems into personal, social settings to corporations and society as a whole.
The things that Business leaders need to be aware of in the rise of machine intelligence and AI in business strategy include cyber security as a major headline to keep an eye on.
Innovation and machine learning and the threat of "modification of facts" to cyber security
We are beginning to connect lots of things, we are beginning to automate lots of things and there is a sub text in the 4th Industrial revolution. What we have seen is we have ideas years and years and sometimes a century before such as Alan Turing's insights into the Thinking Machine; but the step does not happen until the technologies are available to implement the those, and then suddenly society kind of jumps.
I think we are at the point where there are lots of these ideas, and technology has become very cheap and we can implement so many of them including 3D printing, the internet of things, sensors in particular which gives the vast amount of data that machine learning needs. So all of a sudden cheap sensors are enabling machine learning.
So all of these innovations and ideas are coming together, and people are excited about "I can control the heating from my mobile phone". My reaction is, "Wait a minute, where is that information being stored?". The they say "Well it is secure", but that is a grey scale, it is not black and white.
We hear and read about cyber hackers break into things with some much ease what is really secure? This is particularly an issue with Internet of Things security because we are connecting so many things, suddenly. We may think it is wonderful my home knows where I am but so do a lot of other people know.
It is as much about prevention as it is tracking and the impact of machine learning and AI on this such as the recent example of a hackers using ransomware and locking hotel guests put of their hotel rooms remotely demanding a Bitcoin payment to release the system. How do we company executives need to respond to this?
The threat of Modification
Many years ago we developed fingerprint recognition in the late 1970's that by the early 1990's had evolved into an integrated automated fingerprint identification system which later became part of the field of biometrics combining many types of identification. At the time machine driven finger print recognition was wonderful, but nobody was interested even if you "do not need keys anymore" but they could not see it. Ten years later, 911 happened and all of a sudden everyone wanted finger print recognition. Over that period of time all of this biometric information gets stored somewhere as digital information. As soon as that happens it is not the security of it that is the main worry, it can get copied, moved or deleted. Modification is a bigger worry, if somebody gets to the data and changes a link how do you then prove who you are if the original reference data has been edited?
Excerpt ideas draft from the new Book "The 4th Industrial Revolution: An executive guide to applying Artificial Intelligence"
Palgrave macmillan, 2017. Mark Skilton, Felix Hovespian
History of Biometrics January 2015, Biometric update.com http://www.biometricupdate.com/201501/history-of-biometrics
Hotel ransomed by hackers as guests locked out of rooms, The Local, 28 January 2017 http://www.thelocal.at/20170128/hotel-ransomed-by-hackers-as-guests-locked-in-rooms
Andrew Ng shares the astonishing ways deep learning is changing the world , import.io, https://www.import.io/post/andrew-ng-shares-the-astonishing-ways-deep-learning-is-changing-the-world/
What Artificial Intelligence Can and Can't Do Right Now, Andrew Ng, November 2016 , Harvard Business Review, https://hbr.org/2016/11/what-artificial-intelligence-can-and-cant-do-right-now
A Neural Network for Machine Translation, at Production Scale, 27 Sept 2016 https://research.googleblog.com/2016/09/a-neural-network-for-machine.html
Google's AI translation system is approaching human-level accuracy http://www.theverge.com/2016/9/27/13078138/google-translate-ai-machine-learning-gnmt
Microsoft deletes 'teen girl' AI after it became a Hitler-loving sex robot within 24 hours http://www.telegraph.co.uk/technology/2016/03/24/microsofts-teen-girl-ai-turns-into-a-hitler-loving-sex-robot-wit/
AI is now telling doctors how to treat you https://www.wired.com/2014/06/ai-healthcare/
AI disrupting the business of Law https://www.ft.com/content/5d96dd72-83eb-11e6-8897-2359a58ac7a5
Wearable AI can read your emotions as you speak http://www.vocativ.com/398848/wearable-ai-emotions-speaking/
MIT research on reading feelings http://uk.businessinsider.com/mit-radio-waves-emotions-2016-9
Machine generated Art https://deepart.io/