Who is That? AI Voice Manipulation & Leonardo DiCaprio

Much of the deepfake content of the internet is amusing: take this video of Leonardo DiCaprio speaking at the United Nations Climate Summit. “I play fictitious characters often solving fictitious problems”, he says from the podium. However, as we watch Leo, we might laugh as we hear Robert Downey Jr.’s voice boom into the mic. By the end of Leo’s grand speech, we’ve heard from Steve Jobs, Bill Gates, and Kim Kardashian, to name a few.

Deepfake voice manipulation, as showcased in Leo’s speech and executed by ElevenLabs vocal cloning software, is one of the latest abilities and unique products of AI technology that is doing their rounds through Twitter, Instagram, TikTok, and so on. I can see why: wouldn’t it be so cool if I could record a grocery vlog in Morgan Freeman’s voice? However, this same technology carries significant implications for celebrities in the digital media landscape. The ability to manipulate a celebrity’s voice to assume the identities of other individuals, whether they are fellow celebrities, business leaders, or politicians, can take on grave consequences. With voice manipulation, here are a few things that we should be watching out for:

Misinformation and Fake News

Deepfake voice manipulation has the potential to amplify misinformation and the spread of fake news. By convincingly imitating a celebrity’s voice, false statements or fabricated conversations can be created, leading to confusion and misleading the public. Such manipulations can damage the reputation of celebrities and undermine trust in the information disseminated through digital media.

Reputation Damage

Deepfake voice manipulation poses a significant risk to a celebrity’s reputation. Malicious actors can make it appear as if the celebrity endorses certain opinions or engages in inappropriate or controversial conversations. These manipulated recordings can go viral, causing severe reputational harm and potentially affecting personal and professional relationships.

Trust and Authenticity

Authenticity and trust are vital aspects of a celebrity’s relationship with their audience. Deepfake voice manipulation threatens these pillars by blurring the line between what is genuine and what is fabricated. Audiences may become skeptical and find it increasingly challenging to distinguish between real and manipulated content, which can erode trust in celebrities and the digital media they participate in.

Consent and Privacy

Deepfake voice manipulation infringes upon an individual’s right to control their own voice and likeness. Celebrities, like any individuals, should have agency over how their voices are used, especially when it comes to manipulating and distorting their speech for malicious purposes. Unauthorized use of a celebrity’s voice in deepfake manipulations raises serious ethical concerns and highlights the need for legal protections.

Legal and Regulatory Challenges

The rise of deepfake voice technology presents legal and regulatory challenges. Determining liability and establishing frameworks to address the creation and dissemination of manipulated voice recordings can be complex. Legislation needs to adapt to address these challenges, ensuring that appropriate measures are in place to protect individuals, including celebrities, from the potential harm caused by deepfake voice manipulations.

Technological advancements in deepfake detection and verification can aid in identifying manipulated voice recordings. At DeepMedia, we’ve created DeepID, a deepfake detection web app that can identify vocal cloning and audio manipulation, in addition to images and video. We hope this can do good in preserving digital safety. Ultimately, though, addressing the challenges posed by deepfake voice manipulation requires a combination of technological innovation, legal safeguards, and responsible media practices to protect the authenticity, privacy, and trust of all individuals.