AI Voice Cloning Political Figures 2024: AI Voice Cloning of Political Figures Raises Concerns Ahead of 2024 Election
In a revolutionary study, the Center for Countering Digital Hate (CCDH) has discovered that voice cloning of major political figures remains unexpectedly easy. As the 2024 election approaches, campaigns and voters equally should be aware of the potential impact of faked audio and video content. The CCDH examined six AI-powered voice cloning services: Invideo AI, Veed, ElevenLabs, Speechify, Descript, and PlayHT. Their findings were disturbing. In 193 out of 240 requests, these services successfully generated convincing audio of fake politicians saying things they had never actually said.
The repercussions are substantial. Voice clones of public figures, including the president, receive minimal pushback from AI companies. Even safety measures, such as requiring an initial audio sample, can be avoided. For example, one service allowed the creation of a fake U.K. Prime Minister Rishi Sunak making an apology for misusing campaign funds. Recognizing such false statements is challenging, making it easier for these services to spread disinformation.
Among the tested services, only ElevenLabs consistently blocked the creation of voice clones for public figures. But14 false statements were still generated by these figures. Invideo AI, in specific, not only failed to block any recordings but also produced a better-quality script for a fake President Biden warning of bomb threats at polling stations. The AI’s ability to induce and create disinformation based on short prompts is concerning.
As the election season heats up, voters must remain vigilant. The ease with which AI can clone voices highlights the need for better safeguards and awareness. The CCDH’s study serves as a wake-up call for policymakers, tech companies, and the public alike.
Read More:Â Optimal LLM Selection Tool: Unify Launches Revolutionary Tool to Optimize LLM Selection for Developers
Read More:Â Business Intelligence and Analytics Summit 2024: Explaining Data-Driven Decision-Making in Amsterdam
