ARTIFICALLY generated deepfake images and videos could have “life-changing effects” on those depicted, Jersey’s Information Commissioner has said.
The JEP revealed yesterday that a TikTok account showing teachers at Grainville Schools in compromising or humiliating scenarios – all generated through a free software app – had been reported to the police.
The videos were viewed more than 20,000 times as of Friday afternoon, and received over 1,000 shares in total – including people saving these videos to their devices.
Some clips involved teachers getting on all-fours and eating dog food, one teacher holding a gun, and others showing male teachers dancing with the caption “dance if you like little girls” above their heads.
The account – which posed as Grainville School, using their badge, name, and branding it as “official”- was banned from the platform yesterday afternoon.
On Friday, Education Minister Rob Ward said representatives from his department were working closely with staff from the school. Deputy Ward added that it appeared to be an isolated incident, but the use of AI and deepfake technology was being monitored closely across all schools.
Now, Information Commissioner Paul Vane has said the damage of AI-generated content and deepfakes could have “life-changing effects” and stronger deterrence is needed to tackle the growing problem.
He said: “In an island the size of Jersey, reputation can be damaged quite easily, and not for the short-term either.
“The quality of some of these videos is so good that it can be damaging, not just to the individuals, but to society as a whole, because you just don’t know what you’re reading or seeing.
“When individuals are putting together these kind of posts, they’re not really considering the wider implications, either for themselves or for the people that they’re targeting.”
He explained the importance of reinforcing the message that the “responsible use of AI is absolutely key,” and said that those shown in videos could face difficulties when applying for jobs or university.
“Those people who choose to experiment with AI on social platforms may not realise that what feels like a harmless joke can have life-changing consequences for others.
“Deepfakes can depict people in situations they were never part of, causing emotional trauma and reputational damage which can affect them for years to come. In an island the size of Jersey, this effect is amplified considerably”, he added.
Mr Vane added that one of the most difficult challenges in a situation like the Grainville School incident is finding the culprit, as perpetrators often use fake or anonymous profiles, and added that Islanders should consider engaging with an ongoing government consultation on online harms consultation, before the closing deadline on March 6.
He said: “It’s a consultation, so people need to go through the process, read the conversation, understand the comments, and then respond before it closes.”
Building on the responsible use of artificial intelligence, Gus Fraser, a tech entrepreneur, explained that the term AI is a “broad term, covering a wide range of different topics” and that when used correctly, ought to be a powerful tool.
He said: “Not all AI is the same – generating deep fakes is very different from using AI as a study guide.
“It was really easy for me to create some study guides when my daughter was doing her GCSEs, and that was really helpful in terms of her learning.
“I think that sometimes it’s not the services that are being provided, it’s abuse of this technology, and that can be out of the control of service providers. A lot of AI services are provided by Chinese companies, and they’ve obviously got far lower data protection standards and privacy guardrails.”
He explained that despite the easy access of these free platforms, it is primarily down to how they are used.
“This is a behavioural matter, and it’s about changing behaviour. The only way we can do that is through constant reinforcement of the message and further education about responsible use of that kind of technology.”
