Deepfake is one of 13 words singled out for Word of the Year 2023 by the Merriam-Webster dictionary. According to the dictionary, deepfake is “an image or recording that has been convincingly altered and manipulated to misrepresent someone as doing or saying something that was not actually done or said.”

Deepfake: Indian billionaire introduces investment platform, and soccer star Cristiano Ronaldo speaks Vietnamese when selling T-shirts. A deepfake which has gone viral on Facebook among financial investors over the past month is a video showing billionaire Narayana Murthy, co-founder of Infosys, promoting “Quantum AI” as an investment platform that helps generate profits on the stock market. In this video, Murthy is seen saying he and billionaire Elon Musk are jointly developing a project called Quantum AI that is expected to generate huge returns from the stock market. In fact, this is a deepfake stemming from an earlier video about Murthy speaking in July on an economic issue without any reference to artificial intelligence, or AI. Deepfakes can appear in your backyard in Vietnam as well. Recently, a video on social media showing Cristiano Ronaldo promoting blazers, a type of sport jacket, boasting high fluency in Vietnamese. This video is said to be created by AI, copying the face of Cristiano Ronaldo. What if in that video it were not Cristiano Ronaldo but a Vietnamese artist or a certain star falling victim to identity theft? Certainly more people will be scammed. Or what if a deepfake showing Elon Musk pledging to invest in a listed firm in Vietnam? Many will be put in harm’s way if such doctored videos on social media carry the logos of a well-known media house. Deepfake bolstered by social media Previously, it would take a long time to alter a video of Narayana Murthy speaking nothing about AI into a new video promoting Quantum AI. The deepfake technology now enables an amateur to manipulate a video with little time and effort. And, social media networks for quick-view contents such as two-minute clips on TikTok give viewers little time to doubt contents or to verify accuracy. Frauds will be promoted, fake news will spread, and unfathomable damages may occur. One friend of mine working in the communications industry observed that risks are huge for businesses if someone uploads a video on social media about an expert speaking ill of a product or a brand, for instance saying that the product contains cancer-causing substances. The video then can be altered showing that expert was speaking to the mainstream media about the problem product. In the communications landscape, impression and prejudice know no right or wrong. A bad impression about a product will suffice to cause huge damage, my friend said. It has been warned that the year 2024 may witness deepfakes making widespread interventions in major elections worldwide, including elections in the UK and the U.S. Some security and politics experts in the UK have warned that the deepfake technology can be used to create videos misrepresenting what certain politicians speak or do in the run up to the election. Social media platforms will disseminate such videos, and will influence voters when they cast their ballots. Once the prejudice has been imprinted, any efforts to right the wrong could hardly pay off. What the culprits need is to spread fake news with an aim to convince voters, no matter whether the info is right or wrong. The UK’s National Cyber Security Center has warned that new tech apps such as ChatGPT, Text-to-Speech and Text-to-Video tools, or the AI-assisted manipulation of videos can help a person to create thousands of videos from inside his bedroom while previously it took a large team a long time to do so. Therefore, misinformation and disinformation campaigns are forecast to be widespread in 2024. The Labour Party of the UK has been among the first victims. An audio on social media shows Sir Keir Starmer, leader of the Labour Party, abusing his assistants. Mainstream media and the Labour Party immediately stepped in, denouncing the audio as fake, but the clarification was still outpaced by fake news on social media. As many as 1.5 million people have listened to the audio. Sky News remarked that “it is scary we do not know how many people believe the audio is real.” How difficult is it to fake a video? Nguyen Hong Phuc, one of my Facebook friends, says it is as easy as a breeze in a Facebook status on Ronaldo selling sportwear or Singer Phuong My Chi in a Mukbang eating show. “How to fake the voice of a politician? It’s so simple!” he explains. “Simply download AI Coqui-TTS, then look for a five-minute audio recording the voice of that politician, and integrate the audio to AI Coqui-TTS, allowing it to copy the voice. Then you can ask ChatGPT-4 to write a speech according to that politician’s style, and all you have to do is to feed that written speech to AI Coqui-TTS, which will read aloud the speech, right to a T.” Once it is done, simply upload it on social media for dissemination. The big headache is that in certain circumstances, proving a visual or voice recording to be fake is by no means easy, while convincing people to disbelieve what they have seen or heard is all the more difficult. AI and social media have tied the knot to become a well-matched couple able to elevate fraud and shape public opinions. How to cope with deepfake and fake news? First of all, each of us should think twice when uploading anything onto social media, all the more so for influential people since their posts are deemed reliable and often shared by others. Second, mainstream media outlets must make greater efforts to enhance their credibility. I have heard some elderly people in Vietnam say they believe more in what they learn from the Internet than in mainstream media, because they can see “rising prices, business shutdowns and widespread unemployment but newspapers still claim a positive economic outlook.” Given widespread fake news like now, mainstream media must not pursue sensationalism or apply clickbait methods that will erode the public confidence. Once the public confidence in mass media is sapped, the fight against fake news would be more difficult. Third, in my opinion, experts need to actively engage in social media and give the public professional explanations on multiple issues so as to instill the confidence in their professional knowledge. They will become ambassadors of clean information, and important frontliners in the fight against misinformation. The story of a “TikToker being a doctor” offering professional advice to counter misinformation on mental health on social media is a valuable example. Finally, the intervention of competent State agencies is imperative. New regulations need to be issued to assist in the fight against deepfake and misinformation. It looks easier to impose censorship or a ban on communications, but such an approach may backfire as it may cause the public to look for informal information on social media. Humans tend to prefer rumors and concealed secrets, and they also refer to censorship as an excuse to label the fight against fake news as efforts to suppress informal information. Therefore, there should be policies to encourage “white-hat hackers” so as to gather resources in the fight against fake news, promote technologies to authenticate video and voice recordings, and enhance the role of mainstream media and experts in the fight against fake news. The Government needs to accompany other stakeholders in society to suppress fake news via transparency, prompt supervision and assistance rather than via censorship and prohibition.