Eliminating deep fakes from circulation must remain high on the ethical agenda.
If you’re not familiar with the term, they are altered digital images or audio files which make you believe that a cleverly disguised falsehood is a fair representation of the truth. This could be catastrophic if the falsehood were to be acted upon in a global arena.
Sadly, what underlies the deep fake engines are AIs like the Deep Learning Neural Networks at the heart of automated chat messages generated from real voices and video search features which can help to find missing people in the vast pool of unstructured video surveillance data.
However, the ease at which beneficial AI can be put to mischief is alarming, not just to “spoof” politicians but to divert the criminal justice system too.
Stakeholders and courts are struggling to recognise deep fakes and verify that digital evidence has not been tampered with, so much so, it has been reported in the media that deep fakes have already been accepted into proceedings on a number of occasions
The good news is that remedies are on the horizon and not a moment too soon.
For further reading, here is a link to our article “AI, the transparency challenge”
Or visit our downloads page, and access your copy.