In an era dominated by technology, the rise of deepfakes poses a significant threat to individuals, societies, and even governments. Deepfakes, whether in the form of manipulated videos, audio, or images, deceive audiences by making people appear to say or do things they never did.
The US Federal Communications Commission (FCC) recently took action against the use of AI-generated voices after an audio deepfake of President Biden urged citizens to stay home during a robocall in New Hampshire. Similarly, explicit deepfake images of Taylor Swift circulated on social media, drawing widespread concern and highlighting the urgent need for measures to combat this deceptive technology.
The Rise of Deepfakes
Deepfakes are the byproduct of advanced artificial intelligence algorithms that, with just a few taps on a keyboard, can map a person’s face onto another person’s body in explicit videos or images. There’s been a significant tenfold increase in the number of deepfakes detected globally across all industries from 2022 to 2023. There are notable regional differences, namely: 1740% deepfake surge in North America, 1530% in Asia-Pacific, 780% in Europe (including UK), 450% in the Middle East and Africa, and 410% in Latin America.[1]
With AI-driven fraud remaining the most prominent challenge across various industries, crypto is the main target sector (representing 88% of all deepfake cases detected in 2023), followed by fintech (8%).
The Impact
- Pornography: In the domain of nonconsensual pornography, AI-generated content is rapidly escalating, notably impacting minors. It has been found that deepfake pornography makes up 98% of all deepfake videos online, following a 400% increase in deepfake sexual content from 2022 to 2023, reaching monthly traffic exceeding 34 million in 2023, with 99% percent of those targeted being women. [2]
- Misinformation and disinformation: Deepfakes can be used to spread false information or propaganda, which can have serious consequences for public health, political discourse, and national security. For example, during the COVID-19 pandemic, there were several instances of deepfakes being used to spread false information about vaccines and treatments. According to a report by the World Health Organization (WHO), there was a 700% increase in online searches for “coronavirus conspiracy” in the first quarter of 2020 alone.
- As elections loom for half the global population, the widespread creation and dissemination of deepfakes pose a growing threat to democratic processes. The rapid spread of true-to-life deepfakes featuring celebrities and political figures raises concerns about manipulating public opinion. Recently, ahead of Indonesia’s elections, concerns were raised about the technology destabilising the country and manipulating voters. An AI-generated deepfake technology was used to resurrect the late president Suharto, creating a video with his cloned face and voice.
- Frauds and Scams: Deepfakes can be used to impersonate someone, trick victims into revealing personal information or sending money. A 2022 report by Interpol revealed a 300% increase in deepfake-related financial scams.
Steps to Combat Deepfakes
- Legal Frameworks: While the US lacks federal laws do not directly address deepfakes, some states have taken steps. The proposed European Union AI Act also aims to mandate platforms to label deepfakes, establishing a legal foundation for combatting this technology globally.
- Detection Technology: Startups like Sensity AI and Sentinel, along with major tech companies, are developing detection technologies. Intel Corp.’s FakeCatcher claims a 96% accuracy rate in identifying fake videos through
- Digital Watermarks: Companies, including Microsoft Corp., are embedding digital watermarks in AI-generated images to distinguish them as fake. This helps in verifying the authenticity of content.
- Public Awareness: Educating the public about the existence and dangers of deepfakes is crucial. Increased awareness can lead to more cautious consumption of online content and a better understanding of the potential threats posed by deepfakes.
How can policymakers, governments and industry address the challenge of deepfakes?
- Policy and Regulation:
- Clearly define deepfakes within legal frameworks, addressing freedom of expression and potential criminal activities.
- Hold platforms accountable for content moderation, fact-checking, and disclose deepfake usage.
- Promote media literacy through educational initiatives.
- Technological Solutions:
- Invest in research for robust deepfake detection tools for platforms, fact-checkers, and journalists.
- Encourage responsible development of deepfake technology with safeguards like watermarking.
- Standardize metadata to track the origin and modifications of digital content.
- Industry Initiatives:
- Advocate for self-regulation by platforms with stricter content moderation policies.
- Foster fact-checking partnerships to verify content authenticity and flag potential deepfakes.
- Conduct public awareness campaigns to educate users about the dangers of deepfakes.
- Collaboration and Information Sharing:
- Establish multi-stakeholder dialogues for discussing solutions and sharing best practices.
- Promote international cooperation on legal frameworks, technological solutions, and awareness campaigns.
All is not Grim
Deepfake technology presents potential benefits in various domains. One significant advantage is its role in low-cost video campaigns, allowing marketers to create content without the need for in-person actors, resulting in substantial time and cost savings. It also offers a unique way to recreate interactive experiences with deceased artists and celebrities, engaging audiences long after their passing. In the film industry, deepfake has the potential to streamline processes by automating face-swapping, leading to more efficient and cost-effective film production. Additionally, it enhances online learning by creating interactive lecture videos and reconstructing the voices of historical figures. Beyond that, deepfake engages viewers and customers through personalized recommendations, exemplified by Reuters’ use of an AI-generated deepfake for sports news and fashion companies employing virtual fitting rooms for customers to visualize clothing on virtual models with their faces.
Conclusion
In conclusion, while there are benefits, the proliferation of deepfakes presents an alarming threat across various facets of society, from political manipulation and misinformation to the insidious rise of nonconsensual pornography. As governments and organizations grapple with implementing legal frameworks and technological solutions, the responsibility also falls on individuals to protect themselves in this digital landscape. Vigilance, scepticism, and a commitment to digital literacy are paramount. By staying informed, adopting cybersecurity measures, and supporting efforts to combat deepfakes collectively, we can fortify our defences against this deceptive technology. As we navigate the complex challenges posed by deepfakes, a united front of awareness and technological resilience will be essential to mitigate the risks and uphold the integrity of our digital experiences.
[1] https://www.theasset.com/article/50495/asia-pacific-deepfake-incidents-surge
[2] https://www.homesecurityheroes.com/state-of-deepfakes/