Exclusive Networks and Netskope carried out a comprehensive study, focusing on the 15 most widespread fake news generated by AI, taking into account criteria such as social views, engagement, articles, reach and authority to measure the impact of each one.
Fake news trending in 2023
Among the most publicized examples, a video of Hillary Clinton supporting Ron DeSantis, generated by AI, and an image of Pope Francis wearing an oversized white coat stand out, with more than 840 million and 20 million views, respectively. Fake images showing Donald Trump's arrest in Washington DC, created by AI image generator Midjourney, also received significant attention, with over 10 million views and coverage in 671 media posts.
These AI image generators, like DALL-E and Midjourney, make it easy to create fake content, challenging news editors' ability to distinguish the real from the fabricated.
Correction of false stories
The study points out that, on average, a false story about AI takes 6 days to be corrected. Specific cases include an edited video of Bill Gates, which took 15 days to debunk, and the story of an AI drone killing its human operator, which also took a long time to be corrected.
The public's ability to detect fake news
Regarding the ability to detect fake news, more than 80% of participants in the UK study said they were confident in their abilities, with an even higher number in the US. However, when faced with AI-generated fake news, a significant percentage of respondents failed to identify the truth.
Tips for identifying fake news
The study suggests checking the original source, enlarging the image to detect errors and distortions, and paying attention to inconsistencies, such as incorrect proportions or strange backgrounds.
Conclusion
The study Exclusive Networks e Netscope highlights the challenges posed by AI tools in the current era, highlighting the need for media literacy and a critical look at news, especially those generated by AI.
