The global economy is drowning in data. It makes information overload a key concern for many companies. The shortage of talent to process and gain new insights makes it worse.
The answer may be AI. And it was one of the key themes discussed at the recent Enterprise World Asia 2019. OpenText customers, partners, and prospects examined how AI is improving Enterprise Information Management (EIM) in exciting ways.
“As supply chains and IT organizations continue to diversify and outsource, information sprawls and the war for talent intensifies,” said Mark J. Barrenechea, OpenText's chief executive officer and chief technology officer. “Businesses need information solutions and strategies to deliver growth. We are convening the industry to help prepare for this future.”
The ML Detective
Discovery at law firms is one area AI is showing its early promise. Such a process often involves millions of emails and documents.
You need an army of lawyers to process, analyze and ready the documents for the court, said Lalith Subramanian, vice president of Engineering for AI, Security and Discovery at OpenText. He also said the corresponding legal fees would be staggering.
AI can reduce the cost and the man-hours required by finding and cataloging the information.
Security is another area where AI is proving its mettle.
In 2018, OpenText received over two million security alerts. Analysts used AI to filter the massive data to cut through the noise and detect anomalies.
Subramanian also saw a strong link between enterprise information management and AI.
”Because ultimately it’s about managing business risk at the lowest expense and the least risk to you,” said Subramanian. “This is where the concept of enterprise information management comes in and the most highly topical areas for large companies these days are in the areas of compliance, security or legal domain because this is where most of the business risks lie.”
The drawback is that you need the right skills to develop an AI model. It is where OpenText is looking to support.
AI-Augmented Data Analytics
Enter Magellan, OpenText’s AI-augmenting analytics tool that uses open source-based machine learning.
“OpenText Magellan, our AI platform, is designed to understand unstructured content and turn it into quantifiable data,” said Barrenechea.
"It can determine subjectivity, tone and sentiment expressed in the text, nuances that help us derive unique insights from a massive and growing set of social posts, news stories and online comments," he added.
OpenText developed Magellan for the Government of Canada. It helped them to understand media and social opinion on the themes of Canada’s 2018 G7 Presidency. Today, the company markets Magellan as a real-time sentiment analysis tool. It aims to help companies analyze complex, diverse and voluminous data.
It draws on Twitter feeds and online news to create an interactive dashboard. Data analytics identify patterns, relationships and trends to find out the sentiment. It then breaks down the information into critical themes to show what issues are top of mind.
“We’re fortunate that we’re at a point in time where the technology and computing resources are available for the computers to have some sort of cognitive sense of what a piece of text really means,” said Lan Wang, head of Technology, APAC at OpenText. “Using such advanced data analytics to filter out key concepts and analyze the tone of conversations is one way for enterprises to be aware of market sentiment and leads to smarter decisions, faster operations and overall better enterprise performance.”
Think About the Consequences
We are still in the early stages in the use of AI and machine learning in Southeast Asia and the Asia Pacific region -- especially when it comes to enterprise content, said Wang. But while there is huge potential for uptake, Subramanian pointed to enormous challenges on the horizon. Many of these are not technology-related.
“Adoption of AI is also not [solely] a technology problem,“ said Subramanian. “I think if we don’t consider the social impacts, huge as the opportunities are, the challenges for society will also be large. There certainly needs to be deeper, philosophical thought and social leadership that comes up with social structures and support mechanisms for the impacted.”
Developers also need to start looking at the “checks and balances” as they build their AI apps.
“Thinking through the design system in AI models in the early stages and putting in place the right processes with a clear understanding of what the consequences and concerns is absolutely essential,” said Subramanian.
“For example, in the case of autonomous vehicles, if it was on track to hit a pedestrian, what will guide its decision? From a car manufacturer’s standpoint, naturally, the person in the car is the one they are obliged to protect but is that the right thing to do?” he asked.
As with all life-changing technologies, the key questions always center around the philosophical and ethical aspects of how a society should behave. How we answer these will define how AI will shape our future.