Considering the Real Value of Artificial Intelligence In Higher Education

Elon Turner, Director of Infrastructure & Telecommunication Services, University of Arkansas

Elon Turner, Director of Infrastructure & Telecommunication Services, University of Arkansas

Artificial intelligence is a hot topic in information technology, but its real value is hard to quantify and realize. Many technology segments now include AI, machine learning, and data analytics features in cutting edge products. Despite that hype, two areas offer the real potential of applied AI for the education sector that is moving beyond the concept stage: producing insights on student outcomes and managing infrastructure efficiency. Despite many upsides, decision-makers should still bring a healthy level of skepticism to any AI implementation review.

AI is about creating insights from the avalanche of data we are now generating, but many organizations are still struggling with data management. You should start your AI journey with questions like - “what data do we and should we generate,”“how do we securely manage the data?” and “how do we integrate our data.” Your organization needs to have answers in these areas before entertaining an AI that will rely on strong data integrity. Otherwise, you will have a garbage-in, garbage-out effect that will reduce your AI investment value.

Software that evaluates student data is emerging as a real winner for higher-ed organizations looking to improve student outcomes, like retention and graduation rates. Successful applications in this area will apply algorithms, where data on your students go in, and insights are generated. This can be revolutionary, but how do you verify the claims? AI won’t completely take the place of domain knowledge, and poor data can introduce biases into the outcomes. Stay aware of the dangers of blind acceptance of AI as a panacea of insights.

Another area where AI is being applied is to improve the efficiency of infrastructure management like IT operations and campus utilities. These AI platforms for IT operations offer recommendations to solve IT operational challenges. This is built off of the processing of machine-generated data. A human network engineer would previously have to comb through thousands of log messages and have some mental baseline of past performance to arrive at a recommendation. Products like VMware’s Edge Network Intelligence and Juniper Network’s Mavis AI are now to provide senior engineer-level experience and recommendations out of the box by tacking the volume of machine-generated data. These and similar applications can now provide contextual recommendations for managing IT operations better and quicker than your own IT staff. In a challenging environment where it’s difficult to recruit top talent, AI can insert years of experience into your IT team.

New offerings in the facilities AI space are beginning to take advantage of big-data generated by on-premises internet-of-things devices that measure power usage, heating and cooling output, occupancy, and more. Products that analyze this generated information can apply AI to improve utility operations to improve efficiency and reduce costs.

To get the value out of AI, take stock of your organization before committing to vendor claims about their AI. You will ensure that you will be getting the value of cost reductions or improved decision-making. You should spend time speaking to your organization’s stakeholders to understand the level of trust that would exist when deploying AI to ensure that the recommendations can be actualized in your organization. It may be best to find simple, small use cases to get wins with AI that will build trust in your data and trust in the outcomes of the AI.

AI

Information Technology

Data Management

Utilities

VMware

Machine Learning

review