CRF Blog

What do we do when everything online is fake?

by Bill Hayes

In What do we do when everything online is fake? for Chatham House, James Ball discusses the “threat posed by the generation of fake news through artificial intelligence.”

Producing a deepfake video relies on the use of sophisticated AI techniques to make it appear that words spoken by one person are emerging from the face and in the voice of another. This is achieved by building up a bank of words and phrases that the target celebrity has said on film and then using this to extrapolate how they would say the script concerned.

The technique is currently time-consuming, costly and complex, but is expected to become much easier quite quickly, and within a few years could be a feature available on mobile phone apps.

This raises the spectre of a new era of ever more widespread online disinformation, far in excess of the current backdrop of Russian misinformation in eastern Europe, the disruption of American and European elections or the widespread electoral misinformation in Brazil. Imagine what damage could be caused if words were put into the mouth of Benjamin Netanyahu and fake footage broadcast across the Arab world, or into the mouth of Pakistani prime minister Imran Khan and shown across India, or Nigel Farage or Jeremy Corbyn in Britain. [more]