Lorem ipsum dolor sit amet, consectetur adicing elit ut ullamcorper. leo, eget euismod orci. Cum sociis natoque penati bus et magnis dis.Proin gravida nibh vel velit auctor aliquet. Leo, eget euismod orci. Cum sociis natoque penati bus et magnis dis.Proin gravida nibh vel velit auctor aliquet.

  /  Project   /  Blog: 3 Job Listings and how they will be automated

Blog: 3 Job Listings and how they will be automated

There’s no shortage of articles about how jobs will be automated by AI. Just today, this was trending on LinkedIn — 40% of jobs could be automated.

But how exactly will that happen, and what are some of these jobs? It’s one thing to hear about this in abstract, but I think it’s important to consider particular examples. Jobs that are being advertised for today, that will no longer exist in 10 years. Here are a few that I’ve found.

Billing Review Analysts

This job mentions reviewing and auditing medical bills to confirm adherence to regulation. Let’s consider this in 2 parts: reviewing and auditing. To review a bill, the analyst must read several receipts and extract information. Using Optical Character Recognition, a sub-field of Machine Learning, this job could be automated relatively simply (see here for some example OCR solutions). There is even open source/free software out there, like Scikit-Learn!

Now for the next part, auditing. This involves interpreting the information on each receipt (prices) to confirm whether the numbers are in accordance with regulation. Those regulations could be codified into software (or perhaps in a smart contract, in the future), negating the need for a human to do the translation every time.

Let’s estimate that reviewing/auditing a particular receipt takes 2 minutes. When automated, this process would take less than a second.

Legal Assistants

This is perhaps my favorite case on this post. “The daily duties of the IP Docket Clerk will focus on the review and extraction of information from correspondence via email into the docketing software on a daily basis.” Unlike the example above, email is already digital — no need for OCR!

The problem here is referred to as Named Entity Recognition. There are free and paid tools out there to do this already — using this type of software, the job of reading through unstructured emails to extract info will be automated.

Claims Processors

This job involves reviewing claims for workers’ compensation and ensuring necessary documents are present. A checklist can solve the second requirement (one imagines the worker instead uploading document images to the cloud, where an automatic checklist informs on which materials are missing) but for the first, an automatic text parser can easily extract the relevant info — and determine which pieces of data were not present.

Source: Artificial Intelligence on Medium

(Visited 8 times, 1 visits today)
Post a Comment