DeepSummary
The episode discusses the rise of AI-generated deepfakes, which are entirely fabricated audio and video of recognizable people. These deepfakes have become a significant threat to democracy as they can be used to spread disinformation and influence elections around the world. The episode highlights the challenges in detecting and combating deepfakes, as well as the potential consequences of allowing this technology to proliferate unchecked.
Various experts are interviewed, including cybersecurity specialists, journalists, and academics who warn about the dangers of deepfakes and the need for robust solutions to fortify the credibility of critical voices and images. The episode explores proactive and reactive approaches to detecting manipulated media, such as watermarking and cryptographic metadata, as well as the role of social media companies and government regulation in addressing this issue.
The episode also delves into the real-world harms that deepfakes are already causing, particularly to women and marginalized communities. The issue of AI-generated adult content and its implications for consent and online harassment are discussed. The episode underscores the urgent need for action to prevent a future where it becomes increasingly difficult to separate fact from fiction, enabling the spread of disinformation and the erosion of democracy.
Key Episodes Takeaways
- AI-generated deepfakes pose an existential threat to democracy by enabling the spread of disinformation and undermining the ability to discern truth from fiction.
- Proactive measures like watermarking and metadata, as well as reactive measures like detection tools, are needed to combat deepfakes and fortify the credibility of authentic media.
- Deepfakes are already causing real-world harms, particularly to women and marginalized communities, through non-consensual exploitation and online harassment.
- Social media companies have failed to adequately self-regulate, and their business models incentivize the proliferation of harmful content, necessitating government regulation and oversight.
- Robust solutions must involve transparency, accountability, and liability measures throughout the pipeline of AI development and deployment, while respecting privacy and freedom of expression.
- Media literacy and reliance on trusted sources are crucial in navigating the deepfake era, but systemic solutions are necessary to prevent the erosion of democracy and the subjugation of society.
- Combating deepfakes requires a multi-faceted approach involving technology, policy, and education, with a focus on empowering journalists, community leaders, and human rights defenders as the first line of defense.
- The rise of deepfakes underscores the need for collaboration between governments, tech companies, and civil society to develop ethical AI frameworks and safeguard democratic values.
Top Episodes Quotes
- “There are two approaches to detecting manipulated media, what we call proactive and reactive. So the reactive is sort of my bread and butter here. As an academic at UC Berkeley. What we do is we take an image in audio or video, we run it through a battery of tests, and we try to figure out if it's been manipulated or AI generated, all after the fact.“ by Speaker E
- “The onus isn't on the victim to change, it's on the platforms and the government to create tools that prevent these things from happening so easily. If all the women who are being harassed went offline, the trolls would win and this tactic of theirs would be incredibly successful.“ by Speaker F
- “There is no evidence that the technology company can self regulate. The last 25 years have taught us this. And not only that, is that the business model that has led to the Googles and the Facebooks and the tiktoks of the world continues to be the dominant business model of the Internet, which is engagement driven, ad driven, outrage driving. And that business model is the underlying root poison, I would argue.“ by Speaker E
Entities
Company
Product
Organization
Person
Episode Information
Best of the Left - Leftist Perspectives on Progressive Politics, News, Culture, Economics and Democracy
BestOfTheLeft.com
2/28/24
Air Date 2/28/2024
AI-generated deepfakes, entirely fabricated audio and video of recognizable people, are here. They have been on the horizon for years but they have finally arrived during the biggest global election year in history which may prove to be a make of break year for democracy itself as we struggle to separate fact from fiction and autocracy is on the rise around the world.
Be part of the show! Leave us a message or text at 202-999-3991 or email Jay@BestOfTheLeft.com
BestOfTheLeft.com/Support (Members Get Bonus Clips and Shows + No Ads!)
SHOW NOTES
Ch. 1: Deepfaking Democracy: Why AI Threatens News And Global Elections In 2024 - Forbes - Air Date 2-6-24
Deepfaked news segments that appear to be delivered by well-known journalists and TV networks are going viral on social media. Oftentimes, the view counts of deepfaked segments largely outnumber those of real news stories.
Ch. 2: AI Deepfakes Are Everywhere and Congress is Completely Out of Their Depth - CYBER - Air Date 2-9-24
An AI-generated Biden called voters in New Hampshire ahead of the primary and told them to stay home. X locked down the search term “Taylor Swift” after AI-generated nudes of the pop giant flooded the platform.
Tech giants including Microsoft, Meta, Google, Amazon, X, OpenAI and TikTok unveiled an agreement on Friday aimed at mitigating the risk that artificial intelligence will disrupt elections in 2024.
Ch. 4: Deepfake Adult Content Is a Serious and Terrifying Issue - Aperture - Air Date 5-1-23
FINAL COMMENTS
Ch. 12: Final comments on even more dangers from news sites populated with AI-generated content
MUSIC (Blue Dot Sessions)
SHOW IMAGE:
Description: Computer-generated image of a silhouetted man standing in front of a large wall-sized image of the profile of a middle-aged man with digitized lines and cracks running through it.
Credit: "ai-generated-man-computer-science" by NickyPe, Pixabay | License
Produced by Jay! Tomlinson
Visit us at BestOfTheLeft.com