AI Deepfakes Pose Major Threat to Global Elections
With upcoming elections in the United States, as well as in countries like Britain, India, and Mexico, the world is bracing itself for the impact of AI deepfakes on the democratic process. These advanced technologies have the potential to cause widespread disruption, making it increasingly difficult for voters to navigate the landscape of candidate and party policies.
The recent advancements in generative AI, particularly driven by OpenAI’s ChatGPT, have exacerbated the problem. In a disturbing incident, voters in New Hampshire received deepfake robocalls ahead of an unofficial Democrat primary, featuring what seemed to be Joe Biden urging them not to vote. The use of deepfake technology makes it challenging to distinguish between authentic voices and fake ones.
Research published in the journal PLOS ONE revealed that people struggle to detect artificially-generated speech over 25% of the time, highlighting the potential for misuse. These AI-generated deepfakes have also been causing havoc in the UK, where over 100 deepfake video ads impersonating Prime Minister Rishi Sunak were found on Facebook. The paid ads, originating from 23 different countries, including Turkey, Malaysia, the Philippines, and the United States, potentially reached over 400,000 people.
The worrying aspect about the proliferation of AI is that almost anyone with internet access and an AI tool can create deepfakes, raising concerns about the integrity of elections. Ethan Mollick, an associate professor at Wharton, demonstrated this by creating a deepfake video of himself in a few minutes using an AI startup. These developments have prompted AI companies like OpenAI to implement measures to prevent the misuse of AI technology during elections, including the use of guardrails and banning specific tools for political campaigning.
While efforts are being made to combat the spread of AI-generated fakes, such as the work of management consulting firm Oliver Wyman’s social impact team, it remains an uphill battle. Voters are left grappling with the challenge of determining what information they can trust and what has been manipulated by AI. Protecting the integrity of elections requires collaboration from various stakeholders in the democratic process.
In conclusion, the rise of AI deepfakes poses a significant threat to elections worldwide. The ability of AI to create convincing fake voices and videos makes it increasingly challenging for voters to distinguish between real and manipulated content. However, measures are being taken to limit the misuse of AI technology in elections. Ultimately, safeguarding the democratic process will require ongoing efforts and collaboration from all involved parties.
