The French president Emmanuel Macron pledged last week to make France a major international hub of Artificial Intelligence.[i] The French government has committed €1.5 billion over five years to support research in the field, encourage startups, and collect data that can be used safely by organizations and individuals alike.[ii] In ranking the current challenges in moving forward, President Macron underlined that the “first challenge of Artificial Intelligence is that of human talent”. Beyond the politics and economics of the announcement, we can legitimately ask whether AI is proof of innovation, what role does Data Science play here, and how this initiative meshes with the upcoming implementation of the European Data Protection regulations?
What is artificial intelligence?
Artificial Intelligence refers to the ability of information technology to perform tasks commonly associated with human reasoning. The evolution of the Internet of Things has provided a tremendous foundation for the growth of AI — for websites and “apps” are nothing more than platforms designed to offer information “services” in exchange for data on our objectives, thoughts, and actions. Data science fuels the development of AI in identifying pertinent and meaningful information in both small and Big Data sets more effectively. Youyou and his colleagues have demonstrated that inputting 70 « likes » allows a platform to deduce a truer picture of our own characters than that of our best friend, while 150 likes out-performs our own parent or sibling.[iii] The fields of psychometric testing and psychological profiling, popularized recently by Cambridge Analytica, proposes to use this data to predict, if not influence, how humans will interact today and in the future. Is understanding human behavior sufficient proof of innovation?
What do we mean by innovation?
Writing algorithms capable of categorizing and imitating human behavior is one thing, designing machines capable of innovating is another. Although the debate over the nature of innovation is as old as innovation itself, we can concede that the capacity of seeing what others don’t see is one foundation of innovation. Academia research focuses on four applications of innovation: Paradigm, Product, Position, and Process. Researchers have pointed out that innovation itself is largely time-dependent: different generations of innovation (Technology-push, market-pull, strategic integration, and networking ) are defined by the prevailing economic and social factors at any point in time.[iv] Current evaluations of “open innovation” reflect the logic and the justification for both the Experience Economy and the Internet of Things. With this in mind, AI may well be the product of innovation, but it is not a sufficient condition for future innovation.
What keeps managers from seeing innovative solutions to the challenges that defy their organizations and their markets? In decision science, we learn that the major challenges to effective management are the perceptions of the complexity, ambiguity, and uncertainty of the environment in which managers make decisions. In the cognitive sciences, we are taught that our pre-conceptions and prejudices both distort how managers see the problem and bound their ability to propose innovative solutions. In complexity science, we learn of different types of decision environments (simple, complicated and chaordic) that necessitate different forms of inductive, deductive and abduction reasoning.[v] Finally, in business, we learn that innovation doesn’t depend upon our own decisions, but the actions of those around us. As Artificial Intelligence is currently confined to narrow (single task) AI, any discussion of AI’s potential to innovate will depend on the future development of what Nick Bostrom refers to as “super artificial intelligence”[vi] — algorithms capable of reformulating the economic, social and political problems we are trying to solve.
Can AI spur human innovation?
This is an easier question to address affirmatively; let’s use the example of support vector machines (SVMs, also support vector networks) as a case in point. This form of AI uses supervised learning models associated with machine learning algorithms to analyze data in classification and regression analysis. In both cases, the objective here is to optimize the prediction rule (functions) by elucidating the decision boundaries that separate the labeled data. The visibility of these boundaries often requires transforming the space in which the data is represented (from 2 dimensions to 3 or more) to define a hyperplane that allows us to elucidate the linear or non-linear relationships between data points. Leaving the math aside, the takeaway here is that data science allows us to see relationships that are hidden in the familiarity of our own nearsightedness.
In a similar vein, it isn’t much of a stretch to suggest that the future of Data Science lies in our own ability to innovate. Data Science today is still largely bound by the logic of solutions, whereas the roots of innovation lie in uncovering the pervasiveness of apparent problems. Focus differentiates data scientists from the tools of their trade. Whereas analytical packages are designed to identify discernable patterns in the data, a data scientist is trained to focus on the outliers that defy programmable logic. Artificial intelligence isn’t a substitute for human reason, but it can be a fulcrum for our creativity as well as an instigation to explore what AI can’t explain. As Steve Johnson would argue — innovation belongs to the connected mind.
GDPR’s potential impact on AI
A final thought on GDPR’s potential impact on the development of artificial intelligence. As I’ve written elsewhere, Europe’s new General Data Protection Regulation isn’t a direct threat to Data Science, but a legislative plea to encourage data science practices that account for the interests of digital citizens. The development of artificial intelligence can continue to flourish with autotomized data. Moreover, recent technologies seem capable of producing predictive and prescriptive recommendations on “emotional states” without infringing on data privacy laws.[vii] Whether these developments protect individual rights is a separate question and suggests that the future of AI needs to address why we are relying on data science as much as how we are using the data.
Want to learn more?
The practice of business analytics is the heart and soul of the Business Analytics Institute. In our Summer School in Bayonne, as well as in our Master Classes in Europe, the Business Analytics Institute focuses on digital economics, data-driven decision making, machine learning, and visual communications will put analytics to work for you and your organization.
Lee Schlenker is a Professor of Business Analtyics and Digital Transformation, as well as a Principal in the Business Analytics Institute http://baieurope.com. His LinkedIn profile can be viewed at www.linkedin.com/in/leeschlenker. You can follow us on Twitter at https://twitter.com/DSign4Analytics
[i] Thomson, N.,(2018), Emmanuel Macron Talks to Wired about France’s AI Strategy
[ii] AI Technology & Industry Review, (2018), France Pumps €1.5 Billion into AI in Bid to Catch Up
[iii] Youyou, W. et al. (2015), Computer-based personality judgments are more accurate than those made by humans
[iv] Barbeiri et al., (2016), Sixth generation innovation model: description of a success model
[v] Berger and Johnston, (2015), Simple Habits for Complex Times, Stanford University Press
[vi] Bostrom, Nick, (2014), Superintelligence: Paths, Dangers, Strategies, Oxford University Press