The three (f)laws of robotics
Ideas to keep in mind as the AI-ce Age thaws. Artificial Intelligence (AI) has long lived in the recesses of our subconscious with an almost relentless half-life that allows the principle to remain vivid and potent even after countless false dawns over the last 70 years. This is likely to do with our cultural fascination with automation that appears incessantly in various forms of entertainment: from movies, to books, to music. With this we have created a cocoon of fascination where the limits of our knowledge are boundlessly extended to points of contention we are not well-versed in but feel a part of. This is what makes AI such a pressing reality – it concerns us all, but perhaps, the extent to which it will impact us all is not as solidly grasped.
Tweetaway: Ideas to keep in mind as the #AI Age thaws insit.es/2flXReO by Robert Davidson via @InSites #mrx #robotics #artificialintelligence
From 1960s paranoia to Millennial worry
Isaac Asimov famously constructed his Three Laws of Robotics in 1942 which sought to retain an equanimity among robots in his short story Runaround – a part of his on-going composite of robot stories. Looking at these three laws now, we see focus was on safety and not welfare:
- A robot may not injure a human being or, through inaction, allow a human being to come to harm.
- A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
- A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.
Bearing in mind that at this time the atom bomb had not yet been deployed, the purposeful realisation of what machinery could do had not been fully comprehended. Our societal reactions to threats tend to work as a great Rorschach test for our deepest fears, so the idea of mutually assured destruction during the Cold War left a paranoid cloud hovering above a nation worried they would lose everything due to extinction – therefore the notion of a robot sparked a fear of an acceleration of this obliteration. However, much like warfare over the last two decades, we have moved away from pure combat – instead moving to the terrains of economic, political and social disruption. In this new paradigm, at present we should be focusing on the implications for us in terms of what most of us spend most of our time doing: working.
Market research: a truly human venture?
Most thought dedicated to robotics as job-snatchers have zeroed in on the increasing dexterity and sophistication of robots in the manufacturing sector. While the data on these sectors remain murky and a clear indicator of the impact robots will have in this sector is still being thrown together, this sector remains in most minds the obvious low-hanging fruit that an improvement in automation would likely sweep up. We would also throw jobs in the increasingly under-threat sector of transport in there, with an on-going torrent of news which puts self-automated cars on the priority list for this sector, effectively obliterating the need for taxi drivers and other associated positions. This is even one of those sectors which is pitting disruptive technology vs. disruptive technology with some would-be visionaries putting forward the case that no-one will ever need to own a car again as they could just rent one, which is certainly not music to the ears of Daimler AG et al.
We place these professions at high risk because they are in our minds easily replicated, based upon the notion of repetition – something which distinguishes them from white-collar professions such as market research which relies on analysis and in more than common instances, genuine creativity and understanding, something that requires much more than understanding patterns.
This would certainly be the case in the typical vision of AI which looks at robots and automation, but increasingly we are seeing the dissipation of the clouds blocking a true picture of all the activity that is being undergone and its farreaching implications. The greatest motif hidden within this painting is that which runs beneath the surface: machine learning, natural language processing, algorithms and the incoming arrival of general Artificial Intelligence.
The development of general Artificial Intelligence would rocket the intelligence of robots through the roof, with knowledge and comprehension far surpassing our own as an absolute standard.
When looking at this possibility, as well the functionality of cloud computing as a means of further expanding and unifying knowledge, we see the existential threat that Artificial Intelligence poses. While this may seem like a lifetime (maybe even our own) away, current trajectories place the potential arrival of this conservatively at 30-50 years. In-between now and then, we are likely to see the growing role and importance of algorithmbased software to lead the way in reducing the requirement for market researchers such as ourselves. It is this stage that is most prescient and one we should already feel familiar with within our own industry.
Tweetaway: How #ArtificialIntelligence will reduce the requirement for market researchers insit.es/2flXReO by Robert Davidson @InSites #mrx #robotics
What does this mean for market research?
The common argument is that the use of AI (in the short term) will mean that more routine jobs are replaced, meaning focus can be maintained on more creative and exciting work. The loss of jobs to technological advancement is certainly nothing new and dates back centuries with the wheel once being the technological advancement that nullified a large percentage of the work force.
More (relatively) recently, we have seen the advancement of agricultural means drastically reducing the need for human hands, mechanised looms replacing weavers and perhaps most interestingly, drones and other mechanised weapons of war replacing military personnel. Suffice to say, in most cases, this labour has been freed up to pursue more creative means and the argument that technology creates jobs certainly held steady in the IT revolution of the 90s where countless job titles were required to handle the new phenomenon of computers and even more scarily, that deep dark beast – the internet (something InSites Consulting can attest to, utilising the internet to source new forms of market research that have ultimately created our jobs).
However, a stranger shift is taking place currently and that is the huge downsizing of offices for mammoth firms. Facebook, one of the biggest companies (#188 on Forbes list) on the planet, hires just shy of 13,000 employees; compare this to the automotive industry in the 1950s and 1960s in the United States where 1/6 of the American population was employed either directly or indirectly by the automotive industry and you see the stark differences in job creation that previous technological advancements have demonstrated.
Investment banking is to many of us perhaps the most traditional of markets, but as of 2008, efforts have been raised to ensure positive returns and to keep budgets on track and employees are typically some of the highest forms of expenditure for any company – therefore there has been a dramatic downsize in the number of employees without a drastic reduction in costs. It seems the boundaries of what is repetitive work are shifting and this has a lot to do with advancements in Machine Learning.
The capabilities of Machine Learning, software which uses the backlog of data it has performed to help identify patterns, similarities and differences is a big part of this and again, even prior to the arrival of general Artificial Intelligence, this could well start figuring out some of our current tasks. The use of Machine Learning and algorithms could take AI past that key parameter of repetition to understanding. Once this hurdle has been jumped, we can then start to wonder where our jobs might eventually end up.
Already, companies such as Quid have made it their modus operandi to guarantee the role of humans in the future of market research – albeit as a complement to Ariticial Intelligence. For now, this works for Quid as a USP but eventually the USP will be to have it completely analysed by computers – once they have grown capable and affordable, co-creation could see its dying days as a standard.
The changes in the way we are absorbing information is also playing into the hands of our steely friends. The use of infographics and other condensed output is something which AI will become better at through trial and error – eliminating the need for lengthy unread reports which take up precious (and expensive) manhours. While the aforementioned general Artificial Intelligence will attract the thoughts and opinions of the world’s greatest minds such as Stephen Hawking and Elon Musk, generating headlines by itself, this more gradual and unsensational use of AI will largely be ignored. In fact, it has already begun to integrate itself into several other seemingly human-proof industries such as sports reporting. The use of Machine Learning has also been applied to poetry and while the results are terrible at best, the trial and error notion of machine learning combined with human input will help it emerge eventually as a respectable counterpart to emotive, human poetry. The attempts to decode language play an interesting role when thinking of market research. The use of natural language processing which is central to our robot overlords’ deeply disturbing poetry, is also something that market research companies are looking at.
Thoughtly, a firm which looks to enhance research through its Machine Learning services has developed Ellipse, a programme which helps speed up the process of identifying, reading, summarising and visualising huge bulks of text for academic research that could in the future be developed to do analysis on our very own Consumer Consulting Boards (also known as online research communities). Huge leaps and bounds felt elsewhere, most notably IBM’s Watson and the huge steps voice recognition has taken, could mean another system could take on the role of moderator. Of course, this is all unlikely in the short term but certainly food for thought to take into account when thinking of the next 5-15 years of market research.
Will it benefit us?
We do not know what is around the corner, technological advancement thus far has certainly treated us well, it has made the world a safer and more connected place and created countless jobs that we have all directly benefitted from. It could well be that we take the next step and become something similar to Star Child babies in 2001: A Space Odyssey or it could be that we need to deal frequently with psychotic machinery which causes danger – like HAL in Kubrick’s aformentioned masterpiece, or it could of course be both. Such drastic changes won’t happen in the next year but more clouds will begin to float away and we will be able to better view what is hyperbole, what is true and what was unexpected. Understanding where Artificial Intelligence fits into market research may be the last research project we as humans take on.
Tweetaway: Understanding where #AI fits into #mrx may be the last research project we as humans take on insit.es/2flXReO via @InSites #robotics
Stay tuned as next week will be sharing the key take-aways from the ‘Will (y)our next colleague be a robot‘-debate at our Galvanize Congress. More inspiration from the Galvanize Congress (our latest internal congress) is available in our Best of Galvanize Webinar.