‘It’s an indication that AI-generated imagery, especially when it comes to deepfakes, is a global risk that demands a global response,’ said data protection commissioner Brent Homan.
A total of 61 data protection authorities from around the world, including the Crown Dependencies, Spain, Canada, Australia, Mexico, the Netherlands, Hong Kong, Norway and the Philippines, have issued a statement setting out their concerns and calling on all organisations that develop or use AI content generation systems to work with the applicable laws, including those governing data protection and privacy.
The ODPA was involved in helping to coordinate this statement in its role as secretary and co-chair of the International Enforcement Working Group of the Global Privacy Assembly.
A priority of the statement was the concerns of data protection authorities of how AI-generated imagery, including deepfakes could be abused and used in ways that might harm children.
Guernsey is about to see amendments to the sexual offences legislation which will address the creation of deepfakes using AI where there is no consent of the individual who might appear in them. That would make it an offence to create or share intimate images without consent.
Mr Homan said the issue of deepfakes had been around for years, and could be applied to the use of photographic tools to manipulate images.
‘What has been new is the accuracy and the believability. It’s really, really difficult these days to tell a fake image or a fake video from a real one.’
This had led to it becoming easy for someone to create images and to exploit or harass individuals with AI-generated pictures or videos, he said.
In some cases it was clear that the image was satire and Mr Homan cited a recent front cover of The Economist magazine showing US president Donald Trump shirtless and sitting on a Polar bear. But in other cases, things were not so clear, and it is this that the data protection authorities want to see addressed.
Advice has also been given to the public to be careful about what they share online, and checking if it is going to be public or not.
Mr Homan said it was clear that AI was here to stay and that there were many benefits to it.
‘The challenge is how do we harness the power of AI to improve our society while ensuring that it’s minimising and controlling how it can be abused?
‘Because with every great power and innovation throughout history, including the atom, it can be used for good and it can be used for malicious purposes.’
The signatories of the statement said they wanted to encourage the development of ‘innovative and privacy-protected AI’.
They concluded by calling on organisations to be proactive in their engagement with regulators and to bring in robust safeguards from the outset, ‘and ensure that technological advancement does not come at the expense of privacy, dignity, safety, and other fundamental rights – particularly for the most vulnerable of our global society’.
You need to be logged in to comment. If you had an account on our previous site, you can migrate your old account and comment profile to this site by visiting this page and entering the email address for your old account. We'll then send you an email with a link to follow to complete the process.