How machine learning is set to be a major disruptor in the 2020s [Q&A]

Over the last decade we've seen significant advances in AI and machine learning. But there's more on its way, with ML set disrupt almost every industry sector in the decade to come.

We spoke to Eric Loftsgaarden, VP of Data Science and co-founder of consulting services company Atrium to find out how ML can be used to set businesses apart from their competitors and give start ups an edge in traditional industries.

BN: What makes machine learning potentially so disruptive?

EL: Machine learning can be so disruptive because its benefits are so far-reaching. It has the potential to drastically impact every industry over the next decade. For the better part of 20 years, the trend we have seen is analytical reporting for all users. Machine learning is reversing this. Time spent in analytics will start decreasing as machine learning and AI models replace the data interpretation activities and become the analysis engines providing prescriptive actions to business users. This switch from business users having to interpret analytics themselves to being provided prescriptive actions is a massive advantage that can transform businesses.

BN: Will some industry sectors benefit more than others?

EL: As I mentioned, machine learning can disrupt every industry. Industries that may appear less vulnerable to disruption may in fact be affected the most. The effects that Amazon has had on retail and consumer product manufacturing is a good example of this. In my work, I have seen enormous benefits in financial services, healthcare and life sciences, and higher education.

BN: How can it help start ups over more established players, and how can established players stay in the game?

EL: Data-driven startups have the potential to get ahead in industries that do not embrace machine learning. By embracing the emerging technologies, they will have a leg up over competitors. For more established players it is about becoming innovative in collecting and leveraging data their competitors don't have. Companies will be competing to collect more and better data as it replaces other forms of intellectual property as the driver of business and acquisition value.

BN: What are some of the main challenges to ML adoption?

EL: Data quality is a major issue companies face in ML adoption. Your model is only as good as your data. The continued development of data science techniques to better accommodate low-quality data will be crucial in order for companies to better realize the hype around the value of ML and data science. Current expectations around ML are not in alignment with reality.  ML is not just a switch that can be flipped and a complex problem is solved. Effective ML requires the decomposition and re-engineering of a complex process into sub-processes that are better for ML and sub-processes that should remain in the hands of a business user. Another major issue companies face is they don't know what they are looking to do with ML. Successful adoption begins by asking the right questions. What problem am I wanting to solve? If I solve this problem, will it drive growth for my business? Do I have the right data to get started? Do I have a framework in place to implement the insights gained? Companies cannot invest in AI and ML just because it's the big hype train, they have to have a purpose and business goal with it.

BN: What effects will wider ML use have on the jobs market?

EL: With the emergence of ML, people will need to learn how to communicate amongst several different parties within the business. If you want to stay relevant in the job market, you will need to know how to speak across the business. This means knowing the languages of data: math, science, stats, business and IT. Machine learning can provide many benefits to businesses, but that comes at the cost of trusting technology to make important business decisions without fully knowing why. Developing the skills to thrive in an organization where decisions and interdepartmental communication are made based data is key. Machine learning and artificial intelligence are the future of business and I expect the job market to reflect that.

Photo credit: Lightspring / Shutterstock