The term ‘Deepfake’ has come from the term ‘deep learning’. (The process of creating algorithms, where technologies learn from prior experiences.)
Deepfake technology has been around since the 1990s, where it was used in the film industry to manipulate speech. Since the rise of artificial intelligence, we are seeing the rise of deep fakes, and not for good reason.
The boom of the digital era has led modern deepfakes being created to impersonate others. Although many may be deemed as entertaining, when personating a high-power individual, it can come with a risk.
Politicians and celebrities are top of the hit list. Have you seen the infamous Barack Obama deepfake? It’s scarily realistic.
These videos can have a catastrophic impact, if believed, they can alter political results and shift public belief.
With moving image being one of the most powerful pieces of content, you can see why this can be catastrophic. Undermining trust is a huge consequence for the quick mockery videos.
“Very soon, because AI can be more powerful than human editing. You’re not going to be able to distinguish fact from fiction, reality from artificial reality’Adobe
Away from the limelight, in the USA, criminals are using this technology to apply for remote work, the aim to later access company data.
In 2020, the average cost of a deepfake scam was estimated to be $250 million.
Did you know Voice Cloning is also a reality?
Voice Cloning made the news a few months ago after Amazon demoed the possibility of Alexa speaking with a voice of a dead loved one.
The downside of this technology is fraudsters may use it to their advantage. They can easily mimic voices to work in their favour, i.e., pretending to be family members or important authorities. Perhaps fabricated audio will lead the next rise in cyber-crime…
Back in June, a Holocaust educator was able to ‘communicate’ beyond the grave at her funeral. Her pre-recorded information was expertly processed into a deepfake to mimic the natural flow of communication.
In a few years will this become the norm?
Research is currently being conducted on how to determine and counter the threat of deepfakes. As of yet, nothing is conclusive.
– There is an attempt to create deepfake detection technology
– A proposal for blockchain based verification systems
Protection against Deepfakes
For now, there’s a few things we can look out for to protect ourselves against deepfake fraud.
– Lack of, or no blinking
– The skin may be patchy or pixelated
– The lip syncing may be off
– There may be a lack of fine details such as strands of hair
– They may have badly rendered teeth or jewellery
Currently, there is no official ‘deepfake’ law, but the following 3 can be used to support a court case.
– Copyright infringement law
– Defamation legislation
– Anti-harassment laws
Read our last blog post here