Today, AI is not a new concept, particularly among Gen-Zs. However, what is unknown is how AI fraud and deepfakes may occasionally wreck people’s lives, sometimes without them realizing it. Therefore, it is imperative to make sure that AI users truly know what they are using, not just for awareness but also for making the best use of it.

What is deepfake?

Deepfake technology, generated by AI, is not very famous among people in general though it has been a hot topic among tech enthusiasts because it is well known for creating fake personalities, images, videos or audio recordings to impersonate or represent something or someone doing something they didn’t do in real life.

Deepfakes can realistically create content that appears to come from trusted sources. Thus, their ability to spread false information and evidence has a greater impact on people’s lives because even though people know that AI is not human, they tend to trust what is unfolded by AI to be true. There is ample evidence of deepfake usages for notorious reasons such as blackmail, reputational harm, and fraud. This can confuse and harness profiles of celebrities, politicians, entrepreneurs, and even ordinary citizens.

How to spot them?

Spotting “AI or real” is not as easy as people think.

Years ago, when AI was new, images and videos it made were easier to spot because of poor quality and bad lip synching. Studies also discovered that the majority of faces created by deepfakes were not blinking. Yet, technology evolves considerably day by day, and as a result, with its weaknesses getting revealed, they get fixed instantly. So, today, people can find deepfake faces with blinking.

However, when looking closely, people still can find quality-wise issues related to deepfake content where the skin tone seems to be patchy and some fine details such as hair, facial expressions, jewellery, teeth, shadows, and eye movements are visibly inconsistent.

One key tip not to get fooled by deepfakes is the “SIFT” method, introduced by information researcher Mike Caulfield, from the University of Washington. According to this method, you need to,

• Stop

• Investigate the source

• Find better coverage

• Trace claims, quotes and media back to the original context.

People can also use tools such as INVID, DEEPFAKE-O-METER and WINSTONAI to determine if whatever information users obtain from AI is credible and reliable.

People can play games to detect AI and deepfakes to gain analytical and critical thinking on their content because in today’s world, the differences between AI’s ‘real or fake’ are so subtle that they require proper experience and knowledge. This only proves that it will get more and more advanced in the future.

Nonetheless, it is always important to remember that these tools and methods that can spot the differences are there not to stop people from ever using AI but to make sure that they first analyze the content and then responsibly share it.

People might wonder, if deepfakes endanger people this much, why can’t they be banned? The answer is, even when such content has no origin or clarity, it can still be protected by the First Amendment. The First Amendment gives people the “freedom of speech”. The contradictory view regarding this is that deepfake is AI and not real people, and therefore, it does not apply to technology. But this seemingly can again be discredited because just like people’s usage of newspapers, printing presses, photography, radio, and television to distribute speech is justifiable, AI users’ manipulation of information can also be defended.

AI and deepfakes are dangerous, and the speed at which they spread also contributes to this danger but it should not let people completely stay away from using AI because the world is evolving and people need to evolve with it. AI can educate and entertain people, raise awareness and enhance art, music and film industries. So, people should be educated enough to get all these positive and helpful uses of AI.