How Baidu's AI Produces News Videos Using Just a URL

An anonymous reader shares a report: AI for news production is one of the areas that has drawn contrasting opinions. On one hand, it might help media houses produce more news in a better format with minimal effort, on the other, it might take away the human element of journalism or take people out of jobs. In 2018, an AI anchor developed by China's Xinhua news agency made its debut. Earlier this month, the agency released an improved version that mimics human voices and gestures. There's been advancement in AI with text-based news with algorithms writing great headlines. China's search giant Baidu has developed a new AI model called Vidpress that brings video and text together by creating a clip based on articles.

The company has currently deployed Vidpress on its short videos app Haokan and only works with Mandarin language. It claims that the AI algorithm can produce up to 1,000 videos per day, which is a whole lot more than the 300-500 its human editors are currently putting out. Vidpress can create a two-minute 720p video in two and a half minutes, while human editors take an average of 15 minutes to do that task. To train this model, Baidu used thousands of articles online to understand context of a news story. Additionally, the company had to train AI models for voice and video generation separately. However, in the final step, the algorithm syncs both streams for a smooth final video. When you feed the AI algorithm a URL, it automatically fetches all related articles from the internet and creates a summary.

What could possibly go wrong?

[Anonymous Coward]

"When you feed the AI alGorithm a URL, it autOmaticAlly feTches all related articles from the internet and Creates a summary."

What eXtremely bad thing could possibly go wrong?

Re:

[Random361]

Maybe it will create a deadly virus resulting in trillions of dollars in economic damage, hundreds of thousands of deaths, and a global pandemic?

Re:

[michelcolman]

Once it gets smart enough, it can manipulate the news, control what we think and do, and ultimately take over the world. By letting AI serves us the news, we're basically handing it the world on a platter.

damn

[bugs2squash]

more video-only web "pages". That sucks.

Re:

[stabiesoft]

Agreed, I don't know why, but when I want to read a news article, I want to read it not watch it.

So just rehashed data for clicks?

[Stonesand]

Hardly "news". But then again, it was never the intention here to provide "journalism", so maybe this is OK. I guess the headline could better have been: "AI synthesizes available aggregate data into short video form faster than humans can" which is a neat parlor trick. Good job to the devs, though!

Meh

[nospam007]

The summary made me think that an url like www.bloody_car_accident.org would suffice for it to create a video of a bloody car accident.

Not that special

[Malays2 bowman]

So you have a program that fetches the text of a news article, generates a talking head that reads an article with a synthisized voice and lip synchs to what it is reading, and maybe replays some of the video on a site, all in the same bog standard format that TV news channels around the world have been using for decades.

It does not take much thought or AI to put something like this together.

(I didn't need to RTFA, as it's very obvious what they are doing here)

Re:

[Frank Burly]

It makes an interesting dovetail with the article about an Amazon-produced fluff piece that was widely aired because (I suspect) it had good production values and was free.

Re:

[Visarga]

> It does not take much thought or AI to put something like this together.

Not much AI: NLP - classification, clustering, summarisation, text to speech, image synthesis, and syncing them together. Not to mention video understanding for doing cuts from other videos.

This will have higher perceived reliability

[drew_kime]

People are catching on to the fact that you can't trust everything on the internet. Right now, slick video produced in an expensive studio with well-groomed, articulate talking heads is perceived as more reliable. Of course someone is going to try to fake that.

What's the upside?

[t4eXanadu]

I love how the proposed benefit of this is, "more news in a better format with minimal effort".

We already have an endless supply of news produced with minimal effort. Coming soon, even MORE news with even less effort. Yay.

Re:

[algaeman]

This will be ALL NEW low quality CONTENT!

Re:

[wtbman]

All I can imagine is that all of these videos will be complete garbage just like the large number of AI generated Chinese "product review" videos on Youtube.

Re:

[Visarga]

But perhaps in 2 years you will run your own news bot that will make fact checked news reports tailored to your options, thus escaping the flood of fake news.

Give the AI a URL [and] it creates a summary

[grep -v '.*' *]

So give it www.slashdot.org and it'll come up with: "Today, ... ummmm ... never mind."