Email not displaying correctly?
View it in your browser ([link removed]) .
[link removed]
[link removed]
** OPINION
------------------------------------------------------------
Good morning, everyone. Tom Jones is on vacation, but the team at Poynter is keeping tabs on the latest media news and analysis.
** Sora 2 launched to doomsday headlines. How bad is it really?
------------------------------------------------------------
The OpenAI Sora website is displayed on an iPhone on Feb. 16, 2024, in Vancouver, Canada. Sora is a text-to-video artificial intelligence model that generates videos from written prompts. (Koshiro K/Shutterstock)
It’s been almost three weeks since OpenAI launched Sora 2, the latest AI tool that lets anyone generate and share videos of just about anything — with only limited restrictions.
The potential for trouble was obvious from the start. Within hours, users created racist depictions of Martin Luther King Jr., videos of Pikachu being barbecued and SpongeBob posing as Adolf Hitler ([link removed]) .
I was curious about whether Sora 2 has proven to be a greenhouse of deception, as many feared, or if its effects have been more muted so far. So I reached out to PolitiFact reporter Loreben Tuquero and MediaWise director Alex Mahadevan, both members of Poynter’s new AI Innovation Lab ([link removed]) , to hear about their experiences.
Ren LaForme: Headlines about Sora 2’s potential for misdeeds were, in a not very good word, doomsdayish. Are you seeing AI-generated video appear in the fact-checking world, or are we still in more of a brace-for-impact moment? I remember it took a while for AI-generated images to really start fooling people, with the iconic “swagged-out Balenciaga pope ([link removed]) ” being one of the first.
Alex Mahadevan: I can’t speak specifically to what fact-checkers are seeing. But I will say, based on a scrolling experiment I’ve done every day on Instagram since Sora came out, about 25% to 30% of my feed is Sora videos that have been crossposted as Reels. The very first day it came out I got sent three separate ones in earnest. The watermark isn’t super noticeable when you’re scrolling quickly. It happened lightning fast compared to Midjourney coming out and the swagged-out Pope fooling everyone.
Loreben Tuquero: Most of the content I’ve seen is too obviously fake or fantastical to merit fact checks, but it’s not hard to imagine how easily this tool can be used to mislead. For example, some videos show people or objects being sucked into tornadoes, and I worry that bad actors could create similar content to mislead people when natural disasters do strike.
And then, there are the law enforcement bodycam videos that appear to show people being apprehended. With the right prompts, this could lead to reputational damage and sow further division in a fraught political environment. There are ([link removed]) reports ([link removed]) that people are using Sora 2 to create videos of violence against protesters.
Lastly, many videos feature historical figures in modern situations, which people would immediately recognize as fake. But again, if you situate those historical figures in the right context, with prompts that make them say or do something believable enough, those videos might fool a lot of folks. I’ve seen some concerningly realistic depictions of figures like Martin Luther King Jr. and John F. Kennedy.
LaForme: A few days after Sora 2 was released, OpenAI added some guardrails to prevent unauthorized deepfakes. They’ve added some more in recent days. How effective have those been?
Mahadevan: They cracked down on the things that would hit their wallet the hardest, which is the blatant copyright infringement. You won’t see Pikachu doing Pilates or Mario being dragged out of his car by Immigration and Customs Enforcement — or South Park, Dexter or Sopranos characters as much. But, you still see Stephen Hawking getting sucked into a tornado or Kobe Bryant shouting “six-seven” and tackling Bob Barker. Which, to me, is more grotesque and harmful than seeing Cartman doing the robot.
They just started preventing users from creating videos of Martin Luther King Jr., but you can also still easily create videos of fake reporters reporting on fake disasters. Again, way more harmful for the information ecosystem.
So, yes, the guardrails are working, but only on what OpenAI wants them to.
Tuquero: Agreed. Even though users and representatives of public figures can consent to their likenesses being used, that still opens up a breadth of offensive and misleading depictions that can lead to harm.
LaForme: Sora 2 videos come with a watermark, which users have found ways to remove. OpenAI has promised more advanced watermarking. From a fact-checking perspective, what kind of transparency measures could actually make a difference?
Mahadevan: None! There’s too much money to be made with deepfakes and scams, so there are always going to be tools and techniques to remove watermarks or outsmart transparency attempts. Also, it’s near impossible to make a big enough watermark to interrupt someone’s scrolling. That’s my pessimistic view. The trust and safety mechanisms need to be on the production side.
Tuquero: Apart from people finding ways to break guardrails, it’s also becoming even more difficult to develop detection methods and tools at pace with such rapid advancements. One thing OpenAI can do is proactively release its own tool to allow people to identify whether something was made with Sora 2, and invest resources in making it as accurate as it can be.
By Ren LaForme, managing editor
** Trump refiles $15 billion defamation lawsuit against The New York Times
------------------------------------------------------------
After a judge tossed out his first complaint, President Donald Trump refiled his $15 billion defamation lawsuit against The New York Times in federal court Thursday.
Trump first filed his lawsuit against the Times, four Times reporters and Penguin Random House Sept. 15. Just four days later, Judge Steven Merryday, a George H.W. Bush appointee, struck the complaint for being “decidedly improper and impermissible.” At the time, some critics noted that the 85-page complaint read at times like a political rally speech. Merryday found that the complaint violated court rules requiring complaints to be “simple, concise, and direct,” and he gave Trump 28 days to file a new one that was no longer than 40 pages.
The new complaint ([link removed]) is exactly 40 pages and drops one of the Times reporters named in the first one. In the updated complaint, Trump alleges that Times reporters Susanne Craig, Russ Buettner and Peter Baker defamed him in two ([link removed]) articles ([link removed]) and a book ([link removed]) about his rise to power that were published last fall.
“The statements in question wrongly defame and disparage President Trump’s hard-earned professional reputation, which he painstakingly built for decades as a private citizen before becoming President of the United States, including as a successful businessman and as star of the most successful reality television show of all-time— The Apprentice,” the updated complaint reads.
Both the Times and Penguin Random House have defended the journalists’ reporting. Spokespeople for both companies called the lawsuit “meritless,” according to the Times ([link removed]) .
“This is merely an attempt to stifle independent reporting and generate P.R. attention, but The New York Times will not be deterred by intimidation tactics,” a Times spokesperson said.
Trump has filed multiple lawsuits against media outlets over the past two years, drawing criticism that he is attempting to silence and intimidate journalists. Both ABC and CBS settled the lawsuits they were facing, pledging millions of dollars to Trump’s presidential library despite legal experts saying that the law was likely on their side.
In addition to the Times, The Des Moines Register and The Wall Street Journal are facing lawsuits from Trump. He has accused the Register of engaging in “election interference” by publishing an unfavorable poll shortly before the 2024 presidential election and the Journal of defamation for its reporting on a birthday note he allegedly sent child sex offender Jeffrey Epstein. Both the Register and the Journal have denied the accusations.
By Angela Fu, media business reporter
** Pittsburgh Post-Gazette strike passes three-year mark
------------------------------------------------------------
Saturday marked three years since unionized journalists at the Pittsburgh Post-Gazette walked off the job.
It is the longest ongoing strike in the nation, according to the Newspaper Guild of Pittsburgh. Roughly 60 journalists started ([link removed]) striking Oct. 18, 2023, over the company’s refusal to negotiate a new contract. They were also protesting the Post-Gazette’s decision to terminate health insurance for unionized design, production, distribution and advertising staff. (Those staff launched ([link removed]) their own strike Oct. 6, 2023.)
Now, roughly 30 journalists remain on strike, WESA reported ([link removed]) . From the start, the strike was contentious, and the strike vote passed by a slim 38-36 margin. The Post-Gazette continued publishing with the help of the approximately 40 unionized journalists who decided to keep working. Since then, more journalists have crossed the picket line, and the Post-Gazette has hired dozens of replacement workers. Others stopped striking after finding jobs elsewhere.
Within a year ([link removed]) , the union’s strategy became rooted in the court system. With both sides at a stalemate, a court order is likely the only move that will end the strike.
In June, the National Labor Relations Board, which is responsible for enforcing federal labor law, accused ([link removed]) the Post-Gazette of contempt for refusing to comply with an injunction that the company restore the guild’s old health insurance plan and bargain a new contract. The union is currently waiting for the U.S. Third Circuit Court of Appeals to rule on the issue.
Meanwhile, the unions representing the design, production, distribution and advertising staff who had been on strike have since settled, according to WESA. A Teamsters union representing truck drivers settled with the Post-Gazette in April 2024 and dissolved their union, drawing sharp criticism from the other unions then on strike. Then, earlier this year, production and advertising workers took buyouts, leaving just the journalists on strike.
“(A)t the end of the day, what the strike is all about for me is, I don’t want the next generation of journalists to think they have to accept scraps and just scrape by in life to succeed in their career,” photojournalist Alexandra Wimley said in a profile published ([link removed]) by the union’s strike paper.
By Angela Fu, media business reporter
** Media tidbits and interesting links
------------------------------------------------------------
* The Washington Post’s Scott Nover writes “Pentagon press confronts new reality of Trump era: Being banned from the Pentagon.” ([link removed])
* And The New York Times’ Ken Bensinger with “Epoch Times Reporter Resigns After Publication Signs Pentagon Rules.” ([link removed])
* Speaking of … CBS News’ Margaret Brennan spoke plainly about the Pentagon’s new rules ([link removed]) on Sunday’s edition of “Face the Nation.” “Secretary (Pete) Hegseth promised the most transparent administration ever,” she said. “As one former defense official told us, the American people are going to know a lot less about what their Department of War is doing.”
* Also about CBS: Bari Weiss has been the top boss for a week now. How’s it going? The New York Times’ Michael M. Grynbaum and Benjamin Mullin write “Booking Big Guests and Irked by Leaks: Bari Weiss’s First Days at CBS.” ([link removed])
* And more: Variety’s Brian Steinberg writes “CBS News’ Standards Chief Exits Amid Chaotic Period for News Division.” ([link removed])
* For The Washington Post, it’s George F. Will with “What killed print media — and what died with it.” ([link removed])
* The Washington Post’s Kevin Sieff writes “He investigated a soccer star’s death and ended up charged with terrorism.” ([link removed])
** More resources for journalists
------------------------------------------------------------
* Gain the skills to spot AI risks like bias, misinformation and hallucinations before they harm your work. Enroll now ([link removed]) .
* Interested in learning more about funding local news? Start here ([link removed]) .
* Access a list of mental health reporting resources ([link removed]) on funding, source-building and more.
* Stop wasting hours on repetitive tasks — automate them instead. Learn how ([link removed]) .
* Deepen your coverage of incarcerated women and women with incarcerated family members and get the chance to apply for one of five $10,000 reporting grants. Enroll now ([link removed]) .
* Journalists: It’s time to take care of yourself, too ([link removed]) . Enroll now in a free session on embodiment practices to rest and regroup.
Have feedback or a tip? Email Poynter senior media writer Tom Jones at
[email protected] (mailto:
[email protected]) .
The Poynter Report is your daily dive into the world of media, packed with the latest news and insights. Get it delivered to your inbox Monday through Friday by signing up here ([link removed]) . And don’t forget to tune into our biweekly podcast ([link removed]) for even more.
[link removed]
Help Poynter strengthen journalism, truth and democracy. ([link removed])
GIVE NOW ([link removed])
ADVERTISE ([link removed]) // DONATE ([link removed]) // LEARN ([link removed]) // JOBS ([link removed])
Did someone forward you this email? Sign up here. ([link removed])
[link removed] [link removed] [link removed] [link removed] mailto:
[email protected]?subject=Feedback%20for%20Poynter
[link removed]
[link removed]
[link removed]
[link removed]
[link removed]
© All rights reserved Poynter Institute 2025
801 Third Street South, St. Petersburg, FL 33701
If you don't want to receive email updates from Poynter, we understand.
You can change your subscription preferences ([link removed]) or unsubscribe from all Poynter emails ([link removed]) .