The Future of Artificial Intelligence
The Future of Artificial Intelligence
We sat down with Stout's Fotis Konstantinidis to discuss AI adoption, the potential societal impact of new technologies, and the future of AI.
As a brain researcher at the Laboratory of NeuroImaging at UCLA, Fotis Konstantinidis used data collected from brain scans to identify patterns in patients with Alzheimer’s disease. It was during this work that he found his passion for analyzing data and uncovering hidden patterns. At the same time, significant advancements were occurring in the computing space with the onset of cloud computing, which provided increases in storage capacity and computing power at lower costs. With his background in data mining and an education in computer science, Fotis was able to parlay his experience into a career in artificial intelligence, implementing data-driven solutions for a number of Fortune 500 companies across numerous industries. Today, Fotis is a Managing Director at Stout and leads the firm’s Artificial Intelligence (AI) and Digital Transformation practice. We recently spoke with him to get his thoughts on AI trends and opportunities
When people think of AI, they think of robots taking over their jobs. Is this a realistic future with AI? Can you do some level-setting in terms of what AI is and what it can do?
There have been concerns of mass unemployment throughout every major industrial transition (e.g., steam power, mass production, or electronics) in history. By studying these transitions, we find that some workers were displaced due to automation, while others kept their jobs but now performed different tasks due to automation augmenting their previous responsibilities. The overall outcome was the increase of productivity that raised incomes, lowered product prices, and created more jobs than before. I see AI as the driver of another industry revolution, so the effect on the economy should resemble these past transitions. AI is a technology tool that provides accurate prediction and automation to all industries. However, humans are still needed to supervise and provide quality assurance to ensure optimal results. I strongly believe that AI will lead to a significant net job creation and will invent new industry roles that we cannot even imagine today.
What excites you about AI?
What really excites me is leveraging AI to offer data-driven services to solve real-world problems in organizations. I want to move past the hype and the esoteric machine learning algorithms to concrete results that translate into financial and operational benefits for a corporation. AI sometimes feels just like running on a treadmill: It gives you a feeling of gratification, but you always remain stationary, without reaching a destination. I want to ensure that all our AI engagements at Stout have a precise destination, where success is defined based on measurable criteria such as operational cost or customer retention.
What do you view to be the biggest opportunity for AI? What do you view to be the biggest hurdle for adoption?
I believe that AI’s biggest opportunity is to automate and optimize entire industries and launch new business lines that will grow the global economy. Productivity and innovation will thrive, since workers will now focus solely on the strategic and creative tasks. The biggest hurdle for AI adoption, in my opinion, is the workforce reskilling. This is why digital transformations are so difficult. Organizations or governments need to provide the necessary support to their workforce in order to acquire new skills and shift focus to different areas of their job responsibilities. Continuous learning and agility are the key factors in having workers adapt to the new AI-driven era.
Regardless of industry, what are some of the key issues that AI can solve?
Let me start by saying that AI applications have become so prevalent in the last 10 years or so because of the immense computational and storage resources that are available to us after the arrival of the cloud offerings. AI offers the algorithms and techniques that leverage these resources to provide predictive and automating capabilities. AI helps all industries in offering data-driven decisions by analyzing data and simulating all different scenarios and courses of action. Automation of mundane, repetitive or even unsafe tasks is another area where AI becomes a very useful tool. In summary, AI solves industry pain points by automation and prediction systems that provide significant cost savings and increased revenue.
How has AI evolved in the past decade? What new capabilities exist now that did not before?
We now have virtual assistants, such as Apple’s Siri, Amazon’s Alexa, Microsoft’s Cortana, and Google Assistant, that can understand human languages and interact with us in different devices and platforms. Devices accurately capture human motion and complex movements and have become better than ever at recognizing images and playing complex games like chess, poker, or Go. Autonomous vehicles can successfully drive themselves for long distances. AI-based robots are now used in remote or risky areas instead of humans.
How do you think the field of AI will change in the next five to 10 years?
It is really difficult to come up with an accurate prediction for the AI field in the next 5-10 years. Back in the ‘50s, when the field of AI was formally born, the prediction was that we would be able to build a machine as intelligent as a human in 20 years; obviously, this did not happen. I believe that integrating ethics into AI will be an area of focus in the next decade, where ethical and moral reasoning will be used by computers to choose and explain their decisions and actions. We will also see significant advances in the areas of human language comprehension and translation, biomedicine, robotics, and human-machine interfaces and interaction.
Which industries have been the largest adopters of AI?
We see AI being adopted in the automotive industry, specifically in assembly lines and self-driving vehicles, as well as in financial services, in areas such as intelligent fraud detection, anti-money laundering, and loan servicing, to name a few. Marketing, advertising, and retail use AI to understand consumer trends and behaviors. AI is also used in telecommunications and manufacturing for network optimization and predictive maintenance, respectively.
Are there certain industries that can benefit more from AI? Which ones, and why?
My criteria to identify these industries would be a) existence of a reasonable amount of historical data, b) impactful societal benefits, and c) investment amount for modernization. Based on these criteria, I would say that healthcare, education, energy, and agriculture are the main industries that can benefit the most from AI.
Are there certain geographies that have adopted AI more than others? Do you expect that trend to continue?
Since AI adoption requires significant funding from either the public or private sector, the world’s most powerful economies are the ones that have adopted AI more than others. Geographically, countries in North America (U.S., Canada), Europe (U.K., Germany, France), and Asia (China, Japan, India) have been pioneering AI-based products and services. There are also countries with smaller economies, like Israel or Scandinavian countries (Sweden, Norway), where corporations and governments have created the ideal environment for AI to thrive. I expect the same trend to continue, but I do expect developing countries to benefit from the AI advances that developed countries have made.
At Stout, you are responsible for growing the AI & Digital Transformation practice area. Can you tell me more about the services that are offered?
We currently have four major offerings in our AI practice: digital strategy, data/analytics, prediction, and automation. Under our first offering, digital strategy, we assess digital capabilities and platforms and recommend concrete strategies to transform existing digital systems; for example, we provide the build-vs-buy analysis for digital capabilities. Under our second offering, data and analytics, we aggregate and transform data from multiple sources to provide insights and reporting; a sample case study would be ensuring regulatory data compliance. Under our third offering, prediction, we build and train machine learning models to predict certain parameters, such as inventory demand forecast or dynamic pricing. Under our fourth offering, automation, we automate manual, labor-heavy, repetitive processes, such as invoice generation or insurance claims processing.
Can you give me an example of one of the projects you have worked on recently?
We have been working on a variety of projects so far, such as a) application of advanced machine learning algorithms to accurately predict credit ratings of private companies, b) automation of manual processes that identify historical interest rates and bond yields, c) analysis of hundreds of millions of data records to calculate specific financial metrics for companies, and d) processing of patient health-related data to provide useful insights.
How are Stout’s AI services different from what is being offered by other companies?
Our AI practice is comprised of a growing team of talented individuals who are committed to the excellence and quality of service that Stout is known for. We always start by applying design thinking principles to formulate the actual business or IT problems we need to solve. We collaborate closely with our clients throughout our engagements, and our objective is to generate strategies or software assets that are customizable and reusable. Our AI services are not tied to specific platforms or vendors; we always choose the optimal technology tool based on our client’s needs. Most importantly, our primary goal is to offer data-driven solutions that provide real value to organizations, whether this value is reducing costs, increasing revenue, or creating new business lines. Therefore, we always prefer to build quick, working prototypes that demonstrate clearly the benefits of our services.