Nowadays, social media platforms struggle to remove all the fake news content that is posted on their websites and this is a pertinent problem. The concept of fake news is best described as spreading misleading information, which can cause confusion and harm. It can also lead to divided social groups that ridicule each other. An example of this is the COVID-19 pandemic, where false information was spread about the virus and the vaccine that was created to treat the virus. This created groups within society, one that was in favor of the vaccine and one that was strongly against the vaccine, claiming things such as that the government put a chip in it because they want to track you.
For this blog post, I was interested in how generative AI plays a role in the creation of fake news and how easy it might be to create convincing news videos. To do so, I tried to use the tool “Synthesys” to create a very short news flash about Mark Rutte, the Dutch prime minister, and Sigrid Kaag, the Dutch vice prime minister, dating in Amsterdam. I’ve added this short video below and in my opinion, it is evident that this isn’t a real news video and that it is not very convincing either. However, this video was only created in the span of 5 minutes (including creating an account) and given that information, the video actually looks quite okay. It is worth studying how long it would take to really create a convincing news article using various generative AI tools.
An ethical question that the creators of generative AI tools should consider is how they can control the content that is created using their tools. A good example of this is ChatGPT, which I tried to use to create a fake news story or more specifically, a story about Mark Rutte and Sigrid Kaag dating. In both cases, it told me that it could not assist me with that because it involves personal information.
I think this is a really good development and more developers of generative AI tools should consider taking these same measures. If there won’t be any or too little measures, the amount of fake news will exponentially increase and at some point, it will be too hard to distinguish the real news from the fake news. This might result in even more confusion and mistrust in media sources and institutions, which will result in a divided society and political polarization and this should be avoided at all costs.
What do you think about the role of generative AI in the creation of fake news? Let me know in the comments below!
Hi Nivard,
thank you for covering such an interesting topic. Fake news are especially a concerning topic for me as the consequences of some of them can be serious. I agree that developers should have responsible approach when it comes to possible scenarios their tools might create. That is why, I am happy to see creators of ChatGPT covering this issue. Fake news creation reminded me of similar issue in music industry, where AI generated song of Drake and Weekend was streamed by thousands of people online, without both artists knowing about it. To conclude, I am hopeful that further AI implementations will consider legal rights and responsible approaches.