AI in the Spotlight: News Outlets Put Chatbots to the Test

January 27, 2023by TNTMAX

As predicted before the start of the new year, AI is reaching new heights in 2023. 

TNTMAX previously reported how Dell vice chairman and co-chief operating officer Jeff Clarke said that AI would reach an inflection point this year. “2023 is a year that we expect an acceleration of real-world use of AI and machine learning,” he wrote in a blog post. “We have the tools and software systems. We have accumulated significant data sets. We also have invested in training across almost every industry so our teams can graduate from experimentation to real world AI projects.”

In just the first month of the year, there have already been several notable AI updates, specifically in the writing world. BuzzFeed is the latest to jump on the bandwagon, announcing it would begin using ChatGPT, a chatbot tool by creator OpenAI to generate and personalize some of its content and enhance its quizzes. And that news sent shares of the company soaring, rising as much as 150% in afternoon trade.

“To fulfill the promise of our mission, we need to build a stronger business foundation by executing a forward-looking strategy,” BuzzFeed CEO Jonah Peretti wrote to employees.  “We must look ahead and shift our business towards longer term trends in order to seize the opportunities that will come in the eventual recovery,” 

ChatGPT works by using a deep learning technique called “transformer architecture” to sift through several terabytes of data that contain billions of words to create answers to prompts or questions. The tool was reportedly even smart enough to pass prestigious graduate-level exams, though not with particularly high marks.

However, the BuzzFeed news comes as CNET issues corrections for over half of the AI-written articles the outlet recently attributed to its CNET Money team. The articles were not written with ChatGPT, but an “internally designed AI engine.” After the website was first notified of an AI-written article with substantial errors, the company performed an internal audit and CNET Editor-in-Chief Connie Guglielmo says the publication identified additional stories that required correction. 

Guglielmo claims a “small number” needed “substantial correction,” while others had “minor issues” that saw CNET fix things like incomplete company names and language the outlet deemed as vague. Of the 77 articles the publication says were written as part of a trial to test an “internally designed AI engine,” 41 now feature corrections.

With that said, this is gearing up to be the year of AI. Microsoft has even announced a multiyear investment in OpenAI and a person familiar with the deal put its value at $10 billion.