The term “artificial intelligence” dates back to at least the 1950s, and yet, it seems that AI is still in its infancy, given its vast potential use cases and society-changing ceiling.
As AI experts develop a better understanding of both the big data models and applications of artificial intelligence, how can we expect to see the AI market, including machine learning (ML), change? Perhaps more crucially, how can we expect to see other industries transformed as a result of that change?
David Tareen, director for artificial intelligence at SAS, a top AI and analytics company, offered Datamation his insights into the current and future landscape of enterprise AI solutions:
At SAS, Tareen helps clients understand and apply AI and analytics. After 17 years in the IT industry and having been part of the cloud, mobile, and social revolutions in IT, he believes that AI holds the most potential for changing the world around us. In previous roles, Tareen led product and marketing teams at IBM and Lenovo. He has a master’s degree in business administration from the University of North Carolina at Chapel Hill.
Artificial intelligence Q&A
AI as a career
Datamation: How did you first get started in or develop an interest in AI?
Tareen: My first introduction to AI was in a meeting with a European government agency, which wanted to build a very large computer — a supercomputer, so to speak — that could perform a quintillion (1 followed by 18 zeros) calculations per second. I was curious what work this computer would be doing to require such fast performance, and the answers were fascinating to me. That was my first real introduction to AI and the possibilities it could unlock.
Datamation: What are your primary responsibilities in your current role?
Tareen: My primary role at SAS is to improve understanding of AI and analytics and what benefits these technologies can deliver. The AI market segment is noisy, and it is often difficult for clients to understand fact from fiction when it comes to AI. I help our customers understand where AI and analytics can benefit them and exactly how the process will work.
Datamation: What makes SAS a unique place to work?
Tareen: SAS is unlike any other organization. I would say what sets us apart is a deep-seated desire to prove the power of AI and analytics. We are convinced that AI and any of the underlying AI technologies, such as deep learning, conversational AI, computer vision, natural language processing, and others, can have a positive impact on not only our customers and their organizations, but on the world as well. And we are on a mission to showcase these benefits through our capabilities. This relentless and singular focus sets us apart.
More on analytics: Data Analytics Market Review
Understanding AI and key strategies
Datamation: What sets SAS’ AI solutions or vision apart from the competition?
Tareen: There are two areas that make our AI capabilities unique:
First is a focus on the end-to-end process. AI is more than about building machine or deep learning models. It requires data management, modeling, and finally being able to make decisions from those models. Over the years, SAS has tightly integrated these capabilities, so that an organization can go from questions to decisions using AI and analytics.
However, our customers often need more than one analytics method to solve a problem. Composite AI is a new term coined by Gartner that aligns with what we have traditionally called multidisciplinary analytics. These methods range from machine learning, deep learning, computer vision, natural language, forecasting, optimization, and even statistics. Our ability to provide all these methods to our customers helps them solve any challenge with AI and analytics.
Datamation: What do you think makes an AI product or service successful?
Tareen: The key to making an AI product or service successful is to deliver real-world results. In the past, organizations would have little to show for their AI investments because of the hyper-focus on model building and model performance. Today, there is a better understanding that for an AI product or service to be successful, it has to have all the other elements that will help make an outcome better or a process faster or cheaper.
Datamation: What is an affordable/essential AI solution that businesses of all sizes should implement?
Tareen: An absolute must for businesses of any size is a better understanding of their customers. AI is becoming an essential tool to accomplish this. The ability to communicate with a customer the way they like, at the right time and the right place, with the right message and the right offer (as well as making those predictions without compromising data privacy regulations) — that is an essential solution that all businesses, regardless of their size, should implement.
Datamation: How does AI advance data analytics and other big data strategies?
Tareen: With large volumes of data, applying AI to the data itself is a must. AI capabilities can help untangle elements within data, so it can be used to make decisions. For example, we now use AI to recognize information within large data sets and then organize them in accordance with company policy or local regulations. At SAS, we use AI to spot potential privacy issues, lack of diversity, or even errors within big data. Once these issues are identified, they can be managed and then automated, so that new data coming into the database will automatically get the same treatment as it is recognized by AI.
Also read: Artificial Intelligence Market
Trends in AI
Datamation: What do you think are some of the top trends in artificial intelligence right now?
Tareen: In terms of what’s trending in AI, generally there is a lot more maturity when it comes to approaching productive deployments for AI across industries. Gone are the days of investing in building the perfect model. The focus now is on the broader ecosystem needed to deliver AI projects and realize enhanced value. This broader ecosystem includes investing in data management capabilities and deploying and governing AI and analytical assets to ensure they deliver value. Organizations that look at AI beyond just model development will be more productive with their AI initiatives.
Additionally, the notion that AI should be used for unique breakthrough projects has evolved. Now organizations find value in applying AI techniques to established projects to achieve best-in-class results. For example, manufacturers with good quality discipline can save significant costs by applying computer vision to existing processes. Another example is retailers that use machine learning techniques to improve forecasts and save on inventory and product waste costs.
Datamation: What subcategories of artificial intelligence are most widely used, and how are they currently used?
Tareen: AI is really a set of different technologies, such as machine learning, deep learning, computer vision, natural language, and others. All these technologies are finding success in different industries and across different parts of organizations.
Machine learning and deep learning are two areas seeing broadest use with the most promising results. ML can detect patterns in the data and make predictions without being told what to look for. Deep learning does the same but gets better results with bigger and more complex data (e.g., video, images). As these capabilities are being applied to traditional approaches of segmenting, forecasting, customer service, and other areas, organizations find they get better results with AI technologies.
Datamation: What industry (or industries) do you think does a good job of maximizing AI in their operations/products? What do you think they do well?
Tareen: Businesses need to think of AI as more than one technology. Just like people use different senses (e.g., listening, seeing, calculating, imagining) to make decisions, AI can make better decisions when used in a composite way. The most productive organizations combine AI capabilities of computer vision, natural language, optimization, and machine learning into solutions and workflows, which leads to better decisions than their competitors.
Manufacturers are using computer vision to identify quality issues and reduce waste. Banks are having success using conversational AI and natural language processing to improve marketing and sales. Retailers are having success using machine learning in forecasting techniques. As AI gets broader adoption, we should expect to see organizations use a mix of AI capabilities for improved outcomes and different business units and areas.
Datamation: How has the COVID-19 pandemic affected you/your colleagues’/your clients’ approach to artificial intelligence?
Tareen: The pandemic upended expected business trajectories and exposed the weaknesses in machine learning systems dependent on large amounts of representative historical data, including well-bounded and reasonably predictable patterns. As a result, there is a business need to reinforce the analytics “core” and bolster investments in traditional analytics teams and techniques better suited to rapid data discovery and hypothesizing.
As companies adapt to the “new normal,” one of the primary questions we’re asked is how to retrain AI models with a more diverse data set. When COVID hit, the analytical models making good predictions started underperforming. For example, airports use SAS predictive modeling to understand and improve aircraft traffic flow. However, these models had to be retrained and additional data sources added before the models could start accurately predicting the new normal traffic pattern.
More on this topic: How COVID-19 Is Driving Digital Transformation
The future of AI and AI ethics
Datamation: What do you think we’ll see more of in the AI space in the next 5-10 years? What areas will grow the most over the next decade?
Tareen: A complex area where I hope to see growth over the next 5-10 years has large implications for the world: AI algorithms becoming more imaginative. Imagination is something that comes very easily to us humans. For example, a child can see a table as both a table and a hiding place to use when playing a game of hide-and-go-seek. The process of imagination is very complex for an AI algorithm — to learn from one data domain and apply that learning to a different data domain. Transfer learning is a start, however, and as AI gets better at imagination, it will have the potential to better diagnose disease or spot root causes of climate change. I hope this is an area that will grow in the next decade.
Datamation: What does AI equity mean to you? How can more businesses get started in AI development or product use?
Tareen: From inception to now, AI has been used exclusively by subject matter experts like data scientists. Today’s trend is to lessen that need for subject matter experts to instead cascade the benefits of AI to the masses — recognizing the global value from the wide-reaching benefits rather than isolated benefits realized by a select few. The targets for democratized AI include customers, business partners, the sales force, factory workers, application developers, and IT operations professionals, among others.
There are a couple of ways enterprises can push AI to a broader audience: simplify the tools and make them more intuitive. First, conversational AI helps because it makes interacting with AI simpler. You don’t have to build complex models, but you can gain insights from your data by talking with your analytics. The second initiative is to make AI easier to consume by everyone. This means taking your data and algorithms to the cloud to improve accessibility and reduce costs.
Some leaders are surprised to learn that democratizing AI involves more than the process itself. Often culture tweaks or an entire cultural change must accompany the process. Leaders can practice transparency and good communication in their democratization initiatives to address concerns, adjust the pace of change, and successfully complete embedding AI and analytics for everybody’s use.
Datamation: What are some ethical considerations for the market that should be part of AI development?
Tareen: There are numerous ethical considerations that should be part of AI development. These considerations range from data to algorithms to decisions.
For data, it is important to ensure that the data accurately represents the populations for which you are making decisions. For example, a data set should not under-represent genders or exclude low-income populations. Other ethical considerations include preserving privacy and Personal Identifiable Information.
For algorithms, it is important to be able to explain decisions using plain language. A complex neural network may make accurate predictions, but the outcomes must be easily explainable to both data scientists as well as non-technologists. Another consideration is ensuring models are not biased when making predictions.
For decisions, it is important to ensure that controls are in place not only when models are implemented, but that decisions are monitored for transparency and fairness throughout their life cycle.
More on AI and ethics: The Ethics of Artificial Intelligence (AI)
An expert perspective on the market
Datamation: How have you seen AI innovations change since you first started? How have the technologies, services, conversations, and people changed over time?
Tareen: There have been many changes, but one shift has been fundamental. AI used to be overly focused on model building and model performance. Now, there is a realization that to deliver results, the focus must be on other areas as well, such as managing data, making decisions, and governing those decisions. Topics such as bias in data or models are starting to become common in conversations. These are signs of a market that is starting to understand the potential, and challenges, of this technology.
More on data and bias: Addressing Bias in Artificial Intelligence (AI)
Datamation: How do you stay knowledgeable about trends in the market? What resources do you like?
Tareen: My top two places to better understand trends are:
- Spending time with our customers and encouraging them to talk about their projects and possibilities. Our customers in life sciences, manufacturing, health care, banking, and other areas have such good ideas, and those conversations are irreplaceable when it comes to understanding the future of AI.
- Spending time with SAS Research & Development. Visiting Building R on the SAS campus (where the majority of R&D teams are located) and having conversations with engineers as well as the leadership team offers a trove of details that help piece together where technology is heading.
Datamation: How do you like to help or otherwise engage less experienced AI professionals?
Tareen: The key is to describe advanced AI capabilities in ways that are easily relatable and finding examples of customers we have helped in their specific industry.
Work and life
Datamation: What do you like to do in your free time outside of work?
Tareen: One of the benefits of #saslife is work-life balance. I am a private pilot and fly a small aircraft out of Raleigh-Durham International Airport. North Carolina is a pretty state to fly over, so I take as many opportunities as possible to see this beautiful state from the air.
Datamation: If you had to work in any other industry or role, what would it be and why?
Tareen: My ideal role would be one where I can tell real stories about how technologies such as AI and analytics can improve the world around us. Currently, a lot of the work that SAS does, particularly around our Data4Good initiative, fulfills that goal well.
Datamation: What do you consider the best part of your workday or workweek?
Tareen: The interaction with SAS customers is almost always the best part of the workday or workweek. At SAS, we start off every customer meeting with a listening session where we get to hear about their world, their challenges, and what they hope to accomplish. It is an exciting learning process and often the best part of my week.
Datamation: What are you most proud of in your professional/personal life?
Tareen: I am most proud of the work that SAS does around social innovation. Our Data4Good initiative projects are a great way to apply data science, AI, and analytics to big challenges, both at the personal level as well as the global level, to improve the human experience.