What are Deepfakes and why you should know about them?
Deepfakes can make people believe something is true when it is not. It can propagate fake news, help commit fraud and steal millions of dollars. What is scarier is that anyone with a computer and internet can produce deepfakes with tools readily available on the dark web.
So what are Deepfakes?
Deepfakes are manipulated videos and audios. With the use of Deep Learning, malicious minds are able to create fake visuals and sounds to deceive people. They can very convincingly put someone else’s words in another person’s mouth.
Imagine this scenario…a voice message of a CEO of a company is sent to the CFO asking for confidential data or to transfer large sums of money. The CFO who recognizes the CEO’s voice considers it to be the real thing, and goes ahead and unwittingly follows the instructions, causing the company to lose reputation, large sums of money, and more.
These Deepfakes are powerful tools that are extremely sophisticated and very difficult to detect. It is hence imperative to not only know about them but also how to avoid falling prey.
Challenges Deepfakes can create:
From creating social unrest to stealing confidential data, Deepfakes can create chaos and lead to irreparable financial losses.
Such fake media can superimpose inflammatory sentiments on audio, video media to spread malice, incite violence and propagate communal disharmony. They can sway elections by creating fake videos of candidates to influence voters.
Such videos can also cause economic upheaval. A fake video of the Finance Minister can cause the stock market to crash or create panic in investors forcing them to sell their stocks.
If you thought phishing, more specifically social engineering, today is very advanced, Deepfakes are going to take it to an all new level. Cybercriminals will no longer have to hack into corporate systems when they can easily trick employees into giving out sensitive information. The CFO example mentioned above is what social engineering is going to look like in the near future.
According to a Symantec report, three companies have already been victimized in 2019 by Deppfakes of their executives’ voices, losing millions of dollars in the process. This number is going to increase manifold in this decade.
How to detect Deepfakes?
Employees will have to be more vigilant of the audio and videos they encounter and not blindly believe everything they see.
They should take a closer look and check for anomalies. The video could look slightly off. The voice could not completely match. There could be a fuzzy outline around the face. Facial features could be tad different. The ears could look weird and not aligned with the face. If something doesn’t feel right, it is important to flag it off.
However, as Deepfakes evolve and use more sophisticated technologies, telling them apart will get even more difficult. We will see fake videos and audios becoming even more realistic.
The only way forward is to create more awareness about this security challenge and educate people to counter it. Organizations will have to train their employees to become smarter at detecting Deepfakes by catching irregularities.
In the meanwhile, large companies such as Facebook, Google and Microsoft have started their research and are working towards finding Artificial Intelligence-based solutions to help detect and remove manipulated videos.
Businesses will have to find a way to be a step ahead of cybercriminals. They will have to establish a ‘zero trust’ policy where trust is not automatically extended to anyone within or outside the company. Instead, anyone trying to access its systems is verified with credentials. Businesses will have to embed ‘zero trust’ into their DNA if they want to have a fighting chance against Deepfakes.
