China’s Baidu Expands AI Portfolio With Official Release of ERNIE Bot

China’s Baidu Expands AI Portfolio With Official Release of ERNIE Bot

The Gist

  • ERNIE released. Baidu officially launched its large language model, ERNIE bot.
  • Stock boost. Baidu’s stock price surged over 3% following ERNIE bot’s release.
  • Market climb. Baidu’s share of the global search engine market rose to .94%.

With a countrywide ban on Google, China’s top search engine is Baidu. It also ranks among the top five search engines worldwide, claiming .94% of the worldwide market according to Statcounter — up from .65% in February.

Since 2019, Baidu officials have been working on ERNIE bot, as a language model they hope will rival OpenAI’s ChatGPT. And just yesterday, the company shared news of its official release on X, posting in part, “our new-gen generative AI product and large language model, is now fully open to the general public.”

But Baidu was not alone — reportedly on Aug. 15, China rolled out interim regulations to govern the generative AI sector and mandating that service providers undergo security evaluations and obtain approval prior to launching AI products for the mass market. On the heels of receiving this approval, Baidu, along with four other Chinese tech firms launched their artificial intelligence (AI) chatbots.

According to ABC News, following the announcement, Ernie bot went to No. 1 on Apple’s iOS store in China for free apps and the company’s stock price rose by over 3%. Company officials noted the app could be downloaded from app stores or via their official website

In other AI news…

Gannett Benches AI Sports Reporter

While Gannett is a company historically renowned for reporting the news, it’s frequently in the headlines for its seemingly endless rounds of layoffs. But this time the media giant is making news for its decision to halt its AI experiment aimed at automating high school sports coverage after a series of flawed articles put the spotlight back on the timeless debate — machine versus human journalism.

A while back, the company enlisted the services of LedeAI to handle sports briefs — a task once undertaken by humans (including this human when I worked for Gannett back in the day.)

For example, take this AI-written article published earlier this month by the Milwaukee Journal Sentinel. Here are the first two sentences:

The Oregon Panthers defeated the Stoughton Vikings 20-13 in a Wisconsin high school football game on Friday.

Oregon eventually took victory away from Stoughton 20-13 in a Wisconsin high school football matchup on Aug. 18.


Not a venue to let anything slide, X (the social media platform formerly known as Twitter) came alive with reaction.

Sharing a screenshot of a Columbus Dispatch article that also contained a series of repetitive, awkward sentences, Steve Cavendish, editor of the Nashville Banner posted “Yup, that’s a Gannett paper running AI-generated high school football stories. Yup, it’s terrible.”

To which Jon Greenberg, founding editor and senior columnist of The Athletic to reply, “I feel like I was there!”

Yesterday Axios reported that a Gannett spokesperson confirmed “This local AI sports effort is being paused.”

Related Article: Journos vs. Robots: Associated Press Declares Stance Against AI

Penn Study Shows Humans Can Spot the Bot in Byline

A study aimed at distinguishing the difference between human-authored and machine-generated text from the School of Engineering and Applied Science at the University of Pennsylvania revealed that participants were able to detect AI-generated text with an accuracy rate significantly better than mere guesswork.

To gather their data, the researchers used an innovative web-based training game they developed called “Real or Fake Text?”

Source link