Several local SVT editorial offices published AI-manipulated clips of animals after being contacted by a person posing as a tipster.
AI-generated summary
SVT editorial offices were contacted by a person who tipped off about AI-manipulated animal films. The incident was part of a stress test for an upcoming documentary series.
A person who calls himself "Gustav" contacted the local SVT newsrooms in Skåne, Örebro, Västerbotten, Öst, Sörmland, Västmanland and Gävleborg to tip off animals caught on film. Several newsrooms grabbed and interviewed "Gustav" about the animals.
Deepfakes were published
The video clips showed, among other things, a lynx in a garden outside Ystad, a wolverine in a cemetery in Gävle and an otter in Västerås. But the clips of the animals were AI-manipulated and the interview with "Gustav" was a deep fake.
While some newsrooms became suspicious and chose not to publish, the newsrooms in Västmanland, Gävleborg and Skåne fell for the scam.
- There is always a certain human factor when we do news and journalism. What we must do if something goes wrong is to quickly correct it, be open with what has happened and be transparent, and I feel that we were in this case, says Jessica Wennberg, program area manager Sweden, SVT.
SVT series behind the scam
The clip and the interviews with "Gustav" are part of a documentary series that is being recorded right now, where researcher and presenter Gustav Almqvist is doing "field experiments that stress test Swedish society".
- I check what preparedness we actually have to deal with the threats linked to AI. And yesterday I put my team to stress test SVT, he says.
Several media have reported on SVT's blunder. Despite the fact that the method risks damaging SVT's reputation, responsible publisher for the program Axel Björklund thinks that it is justified.
- It is clear that it will be a real stress test for SVT. Sometimes you need to test the routines. Are we equipped for this kind of potential and indeed real risk? says Axel Björklund, publisher responsible for the program.
In the clip, you see how the scam was seen through.

Meta is indefinitely pausing its heavily criticized user agreement that was met with protests from media organizations globally regarding AI training on published material. In Sweden, the Publishers welcome the announcement, even if a permanent solution is lacking.

After the Democrats' success in the mid-term elections, Congress is expected to be able to push through a 'shield law' to protect journalists' sources and increase scrutiny of the FCC under Brendan Carr.

After congressman Andy Harris threatened a journalism student, the debate about freedom of the press and threats to journalists in the United States has intensified. Journalists and experts discuss how the political climate and ownership changes affect news reporting.

A year after journalists were barred from the Pentagon, reporters are testifying about tougher working conditions, reduced transparency and a widespread fear among sources of reprisals, affecting the ability to scrutinize the US military.
SVT has mapped a disinformation campaign on Tiktok where AI-generated videos and images are used to call for protests in Järva. Similar campaigns have previously been detected in Spain, but the sender behind the accounts remains unknown.

Several Swedish newspapers, including SvD, GP, Sydsvenskan, DN and Expressen, are speaking out about their policies around AI use after an essay by Emma Engdahl was depublished due to suspected AI writing.