A written check

For many companies, legal fees from external lawyers can be difficult to track and manage—it can be time consuming and messy to do a line-by-line invoice review—and lawyers can be expensive. One company saw an opportunity to use AI to help businesses track and save on law expenses. After lawyers submit their invoices through this company’s software, artificial intelligence steps in to review them and see if lawyers may have billed for the wrong line items (per billing agreements), and then it flags those line items for human review.
Over time, another AI element of the software also tracks how much different law firms bill you for similar work tasks, helping you identify which firm to chose for certain tasks and even negotiate for better pricing.

What are the implications? This reduces mundane accounting tasks, and it helps reduce lawyer fees.

Photo Found Here: https://www.pexels.com/photo/black-and-white-check-cheque-close-up-259226/

Glasses and computer on a deak

In another life, I reviewed nondisclosure agreements (NDAs) and contracts to ensure the company I worked for knew what it was agreeing to. Back then, this AI-powered NDA review tool would have been a welcomed help. When pitted against 20 able lawyers, this AI had a 94 percent accuracy rate at identifying and labelling different desired elements in NDAs, and the lawyers were accurate on average 85 percent of the time. The decided advantage in this AI’s favor, though, is that it took 26 seconds to complete the task while it took the lawyers between 51 minutes to 2.5 hours.

What are the implications? This tool will augment the lives of lawyers and business workers, saving them time, reducing their job boredom, and allowing them to focus on other urgent tasks.

Photo by rawpixel.com on Unsplash

Slack symbol

If you search for a certain topic in Slack, there’s now a feature that’ll help you identify which one of your accessible coworkers seems to know the most about a topic. This can be really helpful for larger organizations where employees may be spread across the globe.

Slack’ll also soon be rolling out some other features that are accessible upon request, like identifying what employees are talking about, and it can even narrow it down by region.

Photo by Scott Webb on Unsplash

AI added a signature touch to Mercedes-Benz—NLP. You can talk to your car like you would another human: ask it if you can wear sandals tomorrow, and it knows you’re asking about the weather. The system also offers you suggestions through its interface, predicting what you may want to do next based on your user history. Call your spouse? Search for a restaurant? All made easier with this new system.
Note: The video below has background noise because it’s straight from the show floor. If you think that’ll bother you, you may want to skip it.

Photo by Nigel Tadyanehondo on Unsplash

detective with a pencil and magnifying glass

Detectives need to look for similar or related cases when they are trying to solve a crime, and it can take hours of time trying to gather all of the necessary information and make that information easy to understand.

VALCRI, an AI-powered system, can do all of this automatically by analyzing millions of police records and identify patterns that might be useful to the detectives. To try and avoid any bias and erroneous conclusions that the system comes up with, everything is presented to the detectives and they can choose and tell the system what is useful and what isn’t. As a result of this interaction, the AI gets more intelligent as it goes.

Photo by João Silas on Unsplash

Chalk another one up to deep learning from Google. With the tag team of two deep learning networks—one to translate text into a spectrograph (which represents audio frequencies across time) and another to generate speech from that spectrograph—Google can generate voice that is purportedly indistinguishable from a human. If you don’t think you can be fooled, judge for yourself.