Sora 2.0: New AI Generator Creates Fake News in Seconds
amsterdam, woensdag, 8 oktober 2025.
The new AI video generator Sora 2.0 from OpenAI can create fake news segments and radio clips that are almost indistinguishable from real news reports within seconds. Tests by BNR have shown that the system can present incorrect election results as genuine news. Although officially not available in the Netherlands yet, access is possible through a workaround. The generator responds to requests in Dutch, while the same requests in English are blocked, which has serious implications for the reliability of news media and media literacy.
Sora 2.0: A Powerful AI Technology
Sora 2.0, the latest AI video generator from OpenAI, represents a significant step forward in generating realistic video content. This technology can produce complete news segments and radio clips within seconds that are almost indistinguishable from real news reports. Tests by BNR have shown that Sora 2.0 can present incorrect election results as genuine news [1]. This demonstrates the power and potential of the technology, but also the serious implications it has for the reliability of news media and media literacy in the Netherlands [1].
Access and Misuse
Although Sora 2.0 is not officially available in the Netherlands, access is possible through a workaround. BNR’s tests have shown that the generator responds to requests in Dutch, while the same requests in English are blocked [1]. The reason for this is unclear, but it may be that the Dutch version has less stringent security measures [1]. This makes it possible for unscrupulous users to spread false information quickly and easily, which can have serious consequences for the integrity of news media and informed decision-making by citizens [1].
Technical Improvements
Sora 2.0 introduces significant technical improvements over its predecessor, Sora 1. The new version can generate videos up to 10 seconds long with 4K-like detail, compared to the 6-8 second clips of Sora 1 [2]. Additionally, Sora 2.0 better respects physics and object permanence, resulting in more natural movements and fewer visual artifacts [2]. The system also generates synchronised audio, including dialogue, sound effects, and ambient sound, enhancing the immersive experience [2].
Competition and Applications
Sora 2.0 competes with other AI-generated video models such as Google Veo 3, Runway Gen-3, Pika Labs, and Meta’s Vibes. Veo 3, launched in May 2025, has similar features but is limited to 8-second clips, although internal tests have demonstrated clips longer than 1 minute [2]. Runway Gen-3, introduced in mid-2024, focuses on photorealistic outputs and smooth movements, while Pika Labs’ Pika 2.0, launched in the last quarter of 2024, allows users to upload multiple images for scene customisation [2]. Mattel, known for Barbie, is experimenting with Sora 2.0 to speed up the toy design process [3].
Implications for Media Literacy and Democracy
The introduction of Sora 2.0 has serious implications for media literacy and democracy. The ability to generate realistic fake news segments and radio clips increases the risk of spreading false information. This can lead to erosion of trust in news media and reduced informed decision-making by citizens [1]. Disinformation can also influence political campaigns and elections, with serious consequences for the democratic process [1].
Practical Tips for Readers
To identify fake news, it is important to think critically and consult multiple sources. Here are some practical tips for readers:
- Check the Source: Ensure the news comes from reliable and established news organisations.
- Seek Confirmation: Seek confirmation of the news from other reliable sources before sharing it.
- Look for Technical Details: Look for pixelation, unnatural movements, or other technical anomalies that may indicate an AI-generated video.
- Use Fact-Checking Tools: Use fact-checking tools and websites to verify the accuracy of the news.
- Stay Informed: Stay informed about the latest developments in AI technology and the measures being taken to prevent disinformation [1].