BLOG

AI video: will Sora be a gamechanger for campaigners?

By Olly Buston

Whether it’s big budget hero films, or rough and ready clips for socials, video has long been a mainstay of social justice campaigns.  And campaigners have been using AI to generate impactful video content for several years now.  But Sora looks like it could be a game changer.

Way back in 2019 we produced our deepfake videos of Boris Johnson and Jeremy Corbyn endorsing each other to raise the alarm about the damage these emerging technologies could do to democracy.  This was one of the only times we can honestly say we have ‘broken the internet’.  The videos got millions of views and were featured on UK comedy shows, US political chat shows, and even ended up in a German museum.  The process of making our deepfakes was somewhat convoluted as you can see from this excellent BBC ‘Making Of’ video.

Since then these kinds of technologies have got cheaper and more readily available.  But until recently a sizable budget and plenty of tech know-how was still required to create videos like this spine-tingling one from Orange.  And, continuing on a feminist football theme, the campaign for Hope Sogni to be the first female president of notoriously corrupt and misogynistic FIFA.  Another example of a video campaign using AI is the creepy campaign from Deutsche Telekom warning of identity theft.

The pre-launch of Sora from OpenAI suggests that everything is about to accelerate.  OpenAI CEO Sam Altman has been sharing some tantalizing examples of what Sora can do on Twitter/X.  Sora is basically ChatGPT for video.  It turns text into high quality video.  Want to see ‘a video of a bicycle race on ocean with different animals as athletes riding the bicycles with drone camera view?’ Just tap in the prompt and your wish will be granted.

Of course there are major ethical issues to be considered here.  We can sadly assume that the majority of AI-generated videos will be pornographic.  And the Internet Watch Foundation have warned of the increasing use of AI to create child sexual abuse imagery. Meanwhile there are grave concerns about the potential impact of deepfakes on democracy in the year when more voters will vote than ever before.  Add to that issues of bias and plagiarism and you have a very dangerous mix.

We must give serious attention to these profound ethical issues, but at the same time the potential of generative AI video tools like Sora will have creative campaigners licking their lips.  And these tools will surely get more and more powerful quite quickly.  The massive leap in capability from Dall·E 1 (also from OpenAI) to Sora has only taken 3 years.  Who knows what the next 3 years will bring?

 

Email FlorenciaEdgina@FutureAdvocacy.org to secure a place on Future Advocacy’s full day “Full Spectrum AI Advocacy for Beginners” training on Friday 3rd May.

Early Bird Price : £200.  The course is in-person in London.