From science fiction to social media, deepfake technology is not a new concept. So how is it possible to protect yourself against identity theft?
Deepfake technology is not a new concept. Already in 1991, James Cameron featured in Terminator 2, for the first time, an entirely computer-generated character with realistic human movements. Since then, many fictional characters have been created in a fairly realistic and even endearing way, using simple pixels.
This process, which was once incredibly long and expensive, and reserved for experts, has now become widely democratized. In 2021, a TikTok account named @deeptomcruise began posting humorous videos of Hollywood star Tom Cruise. With over 5 million subscribers, this account has transformed into a leading generative AI company, through Metaphysic.ai, making it possible to generate very realistic videos, thanks to generative AI. While this example may make you smile, today, other types of deepfake videos are spreading like wildfire. One of the most recent victims being pop superstar Taylor Swift. While there is no doubt that these questionable videos are not real, the AI-generated images are misleading and are making Internet users increasingly skeptical about the veracity of their origin. While this year, several major events will take place such as the 2024 Olympic and Paralympic Games in Paris or major national elections in the United Kingdom, Europe and the United States, how is it possible to protect against the effects negative of these false contents which usurp our identities?
The realism and risks of modern deepfakes
Today, the vast majority of cyberattacks are identity-related. This means that few companies have actually embarked on a specific identity security journey. However, given the gateway this represents for attacks, they should see this not as an expense but as one of the main investment priorities.
Identity is an essential element of cybersecurity. In business terms, identity means knowing who is accessing what information. In the past, this was usually a person or group of people accessing a database or application. Today, the people who access information within a company are no longer just internal employees, but extend to partners and subcontractors, and sometimes even technological solutions or even artificial intelligence. As for the information available, it passes through more and more systems, mailboxes and entry points; This makes it increasingly difficult to filter identities and secure data against growing threats. Even security measures that were previously thought to be the most advanced and watertight possible, such as voice recognition, are no longer up to the current risks that we encounter with AI.
Identity theft goes very far if we look at the recent example of France 24 and the attempted disinformation of which the media was the victim, just like Emmanuel Macron. This false report claimed that the president had been the target of an assassination attempt during a recent visit to Ukraine. Despite the denials concerning this misinformation, the sharing of this video continues to generate questions and to be relayed. And this is due to the immediate persuasive power of video enabled by this technology, although this is beyond belief. Moreover, as this technology continues to advance, cybercriminals will be increasingly able to abuse it and undermine the security obstacles that stand before them.
The broader, real-world impact of generative AI
It is now against public opinion that the abuses of deepfake are worrying. Over the past decade and a half, it’s clear that the Internet and social media have influenced real-world events, from the use of Facebook in presidential campaigns in the late 2000s to the data analytics scandal personal data from Cambridge Analytics in 2018. But the biggest shake-up lies in the way in which everyday algorithms control how we are exposed to ideas and information. They unconsciously shape the opinion we have on subjects. In the years to come, the risk lies with the growing power of deepfakes and these videos from which it is increasingly difficult to guess whether they are real or not. Will they further improve and influence politics, simply through false information that is ever more difficult to detect?
For several months, deepfakes of political figures making striking, shocking or disconcerting statements have accelerated. Volodymyr Zelenskiy, Donald Trump and Joe Biden are the most notable examples since these videos deceived viewers into believing that they were real. Some may still doubt the nature of these videos and the danger is there that this technology could disrupt the critical sense and confidence of the general public. It could drastically change the trust we have in our political leaders and the credit we can give them. The next legislative elections in the United Kingdom, the European elections and the American presidential election will be tests on the subject. But the same is true for the next Paris 2024 Olympics.
No longer retreat in investing against the enemy
Ultimately, thwarting even a single cyberattack can save millions in potential lost revenue, regulatory fines and reputational damage. However, what is currently holding back investment in identity security is budgetary constraints. And even if in the current context, the savings are understandable, wanting to save a few thousand euros when a real danger is on our doorstep is simply no longer appropriate. IT security teams within companies must have the necessary tools to properly defend against these types of attacks.
Fortunately, even as AI technologies become more accessible, so do security tools. Identity platforms that leverage automation and AI enable businesses to scale identity-related capabilities much faster than businesses without them. Today, investing in tools like this should be the only thing we can take at face value for effective identity protection.
