In part three of this series, we saw how teachers, media companies, and e-commerce businesses are using AI to process large amounts of data and tailor experiences to their end users (or students, in a teacher’s case).
In this post, we’ll be taking a look at how legal and regulatory professionals are using AI to examine risks and improve performance. We’ll also see how those in the science fields are using AI to push their research further.
Risk and Regulatory Compliance
Folks in this space are probably the most cautious when it comes to AI. One thing they are very aware of is that machines, like humans, make mistakes. However, “they could be different from the kinds of mistakes humans make such as those arising from fatigue, anger, emotion, or tunnel vision,” said Vasant Dhar, Professor of Information Systems at New York University.
The stakes are understandably high when these professionals are asked to evaluate state-of-the-art technology. Some current AI applications in this vertical include:
- Performing audits in real time instead of looking backwards at data to find errors.
- Making tedious processes like audits less of a burden (one company reduced an auditing process that historically took a week to “just a few hours.”
Ultimately, the success of AI in the compliance space will depend on the human beings who develop it. There are immense risks involved for professionals who work to reduce risk for their clients or their companies, and giving up control might be too much for some to bear. Additionally, the AI will need to be taught the difference between compliant and non-compliant behavior, which will evolve and change depending on the company’s needs.
According to attorney Mark A. Cohen, “AI is not going to replace lawyers but instead cause lawyers to work differently in the marketplace than they have before.”
Forward-thinking firms are already enlisting AI to take over time-consuming, repetitive tasks that were once the domain of overworked, newly-minted attorneys. Searching records, surfacing old cases, fact-checking, and other data-oriented tasks are being handed over to state-of-the-art artificial intelligence.
Lawyers aren’t generally known to be early adopters. However, a survey in 2016 found that 52 percent of law firms are embracing new technology to do tedious, time-consuming tasks like collecting data, searching records and going through old cases.
Doing grunt work isn’t the only legal application for AI. A Chicago-Kent law professor, Daniel Martin Katz, developed an algorithm to predict decisions in Supreme Court cases. The algorithm was able to predict rulings with a 70 percent accuracy when analyzing 7,700 rulings from a sample spanning 60 years. Another, similar study in the UK found that AI could predict case outcomes with a 96 percent success rate.
Science and Research
Science is nothing if not data-driven, and every generation of scientists has to position its contributions within the context of its predecessors. Biologists are continuing to unravel genomes and log proteins. Astronomers continue to find new star systems and galactic phenomena. And physicists are untangling the strings that bind our universe together (though maybe we’re all just sitting on quantum foam—it’s unclear). All of that information translates to enormous amounts of data and more is being created every day.
In order to continue producing innovative work, scientists have long relied on cutting-edge technology to drive research forward. Today, AI is that cutting-edge technology.
Companies like Iris AI are using machine learning and sophisticated algorithms to help researchers make sense of huge amounts of data within their areas of interest. For example, if you wanted to study the Cicada life cycle in rural Alabama, you could use Iris to map all of the relevant research or papers that would guide your research.
This is possible because AI can translate gigantic data sets into patterns that make sense to human beings. It’s not that we couldn’t achieve the same results given infinite time and coffee—that would just be less efficient. And getting up to speed on what’s already been discovered slows down innovation. What Iris does could save a researcher weeks, if not months, of time. Eventually, Iris intends to develop an AI that could create its own hypothesis, test it, and even publish its findings.
If AI is able to speed up complex tasks, then innovators will have more time to work on innovating.
In our next post (and the last one in this series), we’ll narrow our focus to enterprise and investigate how marketing, sales, and HR teams are using artificial intelligence to support internal growth, as well as customer experience and acquisition.