.jpg)
You may have noticed people on social media talking about something called “deepfakes”, especially in the context of disinformation and the so-called post-truth society. This idea, which became well-known in 2015, talks about a new age of fake news and online information battles. But what exactly are deepfakes, and what risks do they carry?
What are deepfakes?
Deepfakes are fake or manipulated videos (or other content), in which people do or say things that they have never actually done or said, and they can be incredibly realistic. They’re made using machine learning technology, by training a deep learning model on a large amount of data (videos, photos, sound clips) of the person that the deepfake is made of – usually a celebrity or politician. The model can then be used to generate new visual and audial content of this person.
What are the risks of deepfakes?
With advancing development, higher computing power and the availability of AI technologies, the amount of disinformation and false content on social media is increasing. Even worse, most people don’t even know what a deepfake is, much less how to spot them. This leads to an increasing spread of false information, which further undermines trust in real news and information.
Globally, 71% of people surveyed in 2022 knew what deepfakes were. Source: https://www.iproov.com/blog/deepfakes-statistics-solutions-biometric-protection
When people aren’t able to differentiate between real and false information, they become easy to manipulate. Disinformation has already been cited as a threat during important political elections, the COVID-19 crisis and other political and humanitarian crises.
The most common way deepfake technology is used is to create realistic sexual content of specific people. This can include anything from celebrity porn (the majority) to revenge porn. Creating sexual content using a real person’s likeness without their consent can be incredibly damaging to them, and this is also a humanitarian issue.
Deepfake technology can also be used in fraudulent activity – you could receive a phone call that sounds just like the voice of someone you know but is actually someone trying to scam you. An extreme example is a recent incident in Hong Kong: a finance worker, who believed he was in a video conference call with his superiors, who were actually all deepfakes, ended up transferring 25 million dollars to scammers. According to a report by Sumsub, there was a 10x increase in deepfake fraud from 2022-2023.
Deepfakes can also affect individuals' personal lives by creating false and potentially damaging representations of them. For example, a deepfake could be used to create a video of someone engaging in illegal or immoral activities, which could harm their reputation and relationships.
Map showing countries with large increases of deepfake fraud from 2022-2023. Source: https://www.statista.com/chart/31901/countries-per-region-with-biggest-increases-in-deepfake-specific-fraud-cases/
However, a decline in deepfakes is not to be expected in the foreseeable future, quite the opposite. The number of deepfakes has grown exponentially each year since the technology was popularized in 2017. Since the technology is already out there, there’s little we can do to stop people from using it. Instead, we need to become better at knowing how to look out for deepfakes and disinformation in general. Legal measures are also being taken to limit the effects of deepfakes, such as the recently approved EU AI Act, which requires deepfake content to be labelled as such so that people are aware that what they are seeing is fake, although the effectiveness of such measures is not yet certain.
What can you do?
Individuals and companies can take several steps to counter and spot deepfakes. Educate yourself and your employees about the existence and risks of deepfakes.
1. Education and awareness: Implement regular training sessions on how to identify deepfakes by looking for unnatural facial movements, inconsistent lighting, or irregularities in audio quality.
2. Technology Solutions: Use technology solutions that specialize in detecting deepfakes and invest in cybersecurity measures to protect your digital assets.
3. Encourage a culture of skepticism: Encourage a culture of skepticism and verify the source of any suspicious content before sharing it.
By staying informed and vigilant, you can reduce the impact of deepfakes on your personal and professional life. And by standing together against this threat, we can minimize the spread of disinformation and the resulting damage.
