Deepfakes are becoming more common. Here's why you should be worried

KUALA LUMPUR: With a computer and access to the internet, anyone can produce a ''deepfake' content.

Here’s what you need to know.

WHAT IS DEEPFAKE?

• It is a combination of the terms “deep learning” and “fake,” in the form of artificial intelligence (AI).

• A “deepfake” is a falsified piece of content (videos or audio recordings) that has been manipulated or produced using AI.

• According to cybersecurity and defense-focused strategist Peter Singer, the danger of that is the technology can be used to make people believe something is real when it is not.

• Singer adds that deepfakes are extremely convincing and effectively track people into thinking that a person did or said something that never transpired.

• The technology used to produce such digital content has become accessible very easily and quickly.

HOW DOES IT WORK?

• A deep-learning system can create a convincing imitation by analysing photographs and videos from multiple angles of a target person and then copying its speech patterns and behavior.

• All sorts of modification can be done, for example, users can elongate legs, remove pimples and nowadays, some can even generate fake videos that looks very genuine.

• AI devices are already being used to put images of other people’s faces on the bodies of and put words in the mouths.

• Adjunct professor of law at New York University Paul Barrett says a method known as GANs, or generative adversarial networks, makes it more believable.

• The GANs process gets to “identify flaws in the fake content, leading to improvements addressing the flaws.”

According to scientists at the University of San Diego, the uncompressed videos were more than 99 pct successful in deceiving the deepfake detectors in cases where the attacker had access to all the parameters. ETXStudiopic

ARE DEEPFAKES A THREAT TO SECURITY?

• Cybercriminals may use deepfakes in social engineering attacks to fool their targets into giving personal information, bank account details or even money.

• While not all deepfake videos are malicious, the technology does have the potential to cause some real trouble.

• The use of deepfakes to replicate leaders or celebrities may possibly pose a major threat to people’s statuses and to political stability as well.

• Social engineering attacks often relied on impersonation is also an emerging cybersecurity concern.

CAN MANIPULATED CONTENT BE IDENTIFIED?

• While AI can be used to make deepfakes, it can also be used to identify them.

• Many researchers are concentrating on deepfake recognition and looking for a way of controlling it, as the technology is becoming accessible to any computer user.

• But Singer said that identifying the “tells” is becoming tougher and trickier as the deepfake technology develops more advanced techniques, producing videos that look even more realistic.

• The more sophisticated they become, the harder it is to tell that they are fake.

HOW TO PROTECT YOURSELF AND YOUR ORGANISATION FROM DEEPFAKES?

• AI-based detection techniques are being produced that can help us detect fakes, but it’s important to keep in mind that these is never be guaranteed.

• It is vital to always apply verification procedures.

• By enabling privacy restrictions, you can avoid scammers from effortlessly taking your voice, video, or picture.

• One of the main ways to practice good account security is by using multi-factor authentication on every account.