Posted on September 14, 2017
As part of its internship program, Navigator asks its interns to write a blog post about the intersection of communications and an area of personal interest. This week, Pema Banigan examines advancements in artificial intelligence and machine learning and their influence on communications.
By 2020, the market for artificial intelligence is expected to reach $47 billion, and Canada’s contribution to the field will be substantial. Investments in AI have steadily increased around the globe in recent years–a sign that AI is an area with great disruptive potential. The British government has committed $20 million to accelerate its already bustling AI sector; France has organized a national strategy for artificial intelligence; China has pledged more than $2 billion towards AI; and the U.S. has budgeted $175 million in direct funding.
Canada is taking steps to support its budding AI industry as well. In Budget 2017, the Government of Canada announced it intended to renew and enhance its funding for the Canadian Institute for Advanced Research (CIFAR), investing $35 million over the next five years. CIFAR’s mission is to enable transformative knowledge by connecting the world’s pre-eminent researchers, and to fund the pioneers of deep-learning AI. The government also announced that CIFAR will administer a $125 million Pan-Canadian Artificial Intelligence Strategy for research and talent development. The Pan-Canadian Artificial Intelligence Strategy intends to attract and retain top academic talent in Canada, increase the number of postgraduate trainees and researchers studying artificial intelligence, and promote collaboration between Canada’s main centers of expertise in Montreal, Toronto-Waterloo and Edmonton. Canada is in the AI race.
With huge investments and advancements in technology, those of us not in the field still have a vital role to play in the adoption and application of these emerging algorithms. How we integrate new programs and data insights into our businesses will shape the landscape of emerging AI companies in Canada and around the world. As with all new technologies, invention and engineering is half of the story; the ways in which consumers employ the innovations can be just as creative. The process is reciprocal. At Navigator, our clients depend on us to guide them in understanding, and responding to, public opinion. Research is an integral pillar of our business and we anticipate AI will yield significant opportunities to gather data more intricately than ever before.
While we may aspire to Homo Economicus – the ideal human who possesses the infinite ability to make rational decisions – a rational person knows we will always be affected by conscious and unconscious biases. Biases and heuristics are an inescapable part of the human cognitive process. As Nobel laureate Herbert Simon would say, our rationality is ‘bound’ – limited by the tractability of the decision problem, the cognitive limitation of our minds, and the time available to make the decision. Simon dismantled the traditional economic assumption of the rational-decision maker, arguing that humans have a limited capability to process vast amounts of information available to them. The speed and scale at which AI programs can pore over data is its great strength.
Public opinion and market research AI programs that rely on ‘sentiment analysis’ should be of great interest to all business leaders and decision makers. Sentiment analysis – also referred to as Opinion Mining or Emotion AI – sits on top of natural language processing. With a multitude of possible algorithms – for example Naive-Bayes, where words are sorted into categories and allocated a positive or negative number – subjective information about affective states can be extrapolated en masse. When done skillfully, the attitudes and opinions of an entire population can be analyzed and demographic differences extrapolated. Social media data and public information of a massive representative sample, for example, can be sifted through overnight with exponentially great power of speed and scale. The impact technologies such as these could have on our decision-making process is therefore tremendous – offering a clearer lens with which to analyze data.
The assertion that our ability to process data is limited is taken further when considering the work of behavioural economists Daniel Kahneman and Amos Tversky. They identified a number of cognitive biases and heuristics that people rely upon when making decisions, such as the anchoring effect (a common bias where people rely too heavily on the first piece of information they are exposed to), the availability heuristic (a shortcut where greater probability is assigned to the first example that comes to mind), and confirmation bias (the tendency to interpret information in a way that confirms existing beliefs). In general, heuristics are quite useful to reduce the complex task of assessing probabilities into simpler judgmental operations – but sometimes they lead to severe and systematic errors.
Most people suffer from overconfidence in noisy environments when the feedback is weak. Businesses–and those who run them–constantly have to make decisions in such circumstances. Data insights from AI programs offer a pristine perspective on the world, helping people overcome their biases. With the advent of the ‘internet of things,’ more and more data is becoming available, strengthening the programs and offering greater external validity. The partnership of big data insights and human expertise will allow for greater accuracy and efficiency in the future.
The lesson is clear: AI can open up numerous opportunities for analyzing big data and understanding human sentiment. As we approach the next stage of AI, all organizations should reflect on how they can integrate advancements into their internal processes. It is important to remember that, although these big data insights may offer a new lens, the right insights can only emerge with intelligent and experienced human judgment and analysis.