HeyGen: what it is and what risks it poses

At least once in our lives, we've all wished for a look-alike to step in for us in an exam or work. What if we told you that this could actually happen for many people working online? We have HeyGen, the latest artificial intelligence software, to thank for this possibility. This toll it's causing a sensation on social media. Several creators, like the TikToker Vickybeauty, have decided to bring their digital doppelgänger to life, capable of replacing them in a video content. What makes this simulator so astonishing compared to the ones we hear about every day? The avatar created by HeyGen not only faithfully reproduces the features of the user but can also speak many languages.

A possible application and a meaningful reflection

Giving a presentation in German or French won't be difficult thanks to this AI software, as the lip movements sync perfectly with the words, and it all looks incredibly natural. To use it, you just need to upload your photos and videos to the website and write the text you want your alter ego to say in the appropriate section. Besides expressing a preference for a certain language, you can also record audios so that the AI can replicate your voice. "The weird thing is that I couldn't tell the real me from the artificial intelligence, and this could be a problem in the future" observed Vicky after using the app. Her perspective is very useful for sparking a public debate on this topic.

Celebrities selling their features to Meta

These days, we're witnessing an unprecedented phenomenon involving influencers and celebrities. Charlie D’Amelio, Kendall Jenner and others have sold their image to Meta. As a result, they will be transformed into chatbots capable of interacting with users on Instagram, Messenger, and Whatsapp. Voice and appearance are the only things these virtual assistants have in common with the identities of the VIPs. Their names have actually been changed to shorter, easier-to-remember ones like Coco and Billie. In the profiles of these virtual entities, you can scroll through posts and reels, just like in a regular account. However, the photos you see are entirely generated by AI.

Zuckerberg's words

According to Zuckerberg, offering such a service can serve a dual purpose, satisfying both the producer and the consumer. In the future, companies can take advantage of these open-source technologies to automate and personalize daily interactions with consumers. This is thanks to Meta AI Studio's new program. In Zuckerberg's words: "Other companies can also create AI characters that reflect their brand's values and enhance customer service experiences. From small businesses looking to scale their activities to large companies looking to improve communications, AI can help businesses engage with their customers through our apps." However, this approach, which will forever transform digital interaction, customer service, e-commerce, and content distribution, is not without its downsides.

The privacy issue

Among the many concerns raised by users, the risk of jeopardizing one's privacy is certainly at the forefront. "It's important to know that we train and optimize our generative artificial intelligence models to limit the possibility that private information a user might share with AI tools appears in responses to other people. To examine interactions with our artificial intelligence, we use both automated technology and individuals, in order to, among other things, reduce the likelihood that model outputs include someone's personal information and improve model performance." With these words, Zuckerberg wanted to reassure digital citizens alla around the world. However, in Italy, the function to start a conversation with one of the 28 avatars created by Meta is not yet available.

Online ethics

Now more than ever we need to clarify the intentions of professional codes, international regulations, and online ethics. Several experts in ethical philosophy and contemporary sociologists have pointed out the risks involved in all of this. Since 2021, the EU has introduced the concept of artificial intelligence risk, and has created a scale to measure it, ranging from minimal to unacceptable. In addition to the proposal for a regulation on the European approach to artificial intelligence, more recent bills have been introduced, such as the one presented by Senator Antonio Nicita, on the transparency of artificial intelligence. This proposal involves introducing the obligation to provide, for subjects who distribute AI-generated content, a unique label that allows automatically generated content to be immediately distinguishable from content generated by humans.

The sinister uses of AI

It's true, technologies like HeyGen could help you cover up that little lie you put on your resume with your employer, send a birthday message to your friend whose birthday you always forget before midnight, and always appear dressed up in work calls. But what if someone used the same tool to commit a crime, or perhaps defame someone on social media with your image without your consent? If some tech giants have decided to sign the famous open letter "to pause the advancement of artificial intelligence" then perhaps we should also stop and reflect. The debate is still open: there are those like Meta who promise absolute respect for Privacy regulations, and there are those who do not trust the promises made in the letter, like Elon Musk and his new startup "xAI." What awaits us? We'll have to wait and see.