This post was featured in the Reality Defender Newsletter. To receive news, updates, and more on deepfakes and Generative AI in your inbox, subscribe to the Reality Defender Newsletter today.

Deepfake of the Week

Never mind the 2024 elections. Last week saw Toronto mayoral candidate Anthony Furey use generative AI in official campaign materials to push a policy issue. 

While speaking on combatting homelessness in the city, Furey’s materials depict homeless encampments and rows of idealized text that appear real at first glance, but were wholly generated upon our analysis.

What does the Reality Defender Platform say?

What does the Reality Defender team say?

Deepfakes and AI-generated media are now the norm in every election cycle. Our team has yet to see any instance of AI-generated video or images used in promotional materials by political candidates that are explicitly labeled as such. As long as legislation fails to emerge in every country where this is a problem (which will soon be every country), candidates will continue to use synthetic media in an effort to deceive or defame without any indication of said media being inauthentic.

The effects of deepfakes and AI-generated media being constantly used and increasingly commonplace in the political sphere are too numerous to name. Countering them after their widespread release and subsequent virality is akin to trying to put the toothpaste back into the tube. 

Reality Defender’s proactive deepfake detection can be deployed today to grind this type of disinformation to a screeching halt. Our clients, which run the gamut from large enterprises to government agencies, are using Reality Defender to solve their immediate deepfake-driven problems at this very moment. Yet when defamatory images created using generative AI are convincing enough to sway large swaths of voters, implementation of measures to combat them should be among every elected official’s highest priority — before they find themselves on the receiving end of AI-generated disinformation.

(Deepfake) Lennon–McCartney

Ex-Beatle Paul McCartney revealed last week that the long-deceased John Lennon’s vocals will be generated by AI to finish a never-before-released Beatles song. Though AI-assisted music is increasingly becoming more prevalent, and the Beatles were early adopters of technology that is now widely used in rock music (including, but not limited to, artificial double tracking, Moog synthesizers, and sampling), this is the most famous instance yet of recreating a deceased person’s voice and using it in a popular work.

Arizona Woman Raises Alarm Over Voice Deepfakes

Jennifer DeStefano recounted her traumatic experience of a scam using artificial intelligence  to mimic her daughter's voice to a Senate judiciary committee last week, highlighting the nefarious potential of AI in criminal hands. DeStefano urged lawmakers to regulate the technology, stressing that if left uncontrolled, these kinds of scams could undermine public trust and our understanding of truth. This was further evidenced by a recent McAfee survey which found 70% of respondents couldn't confidently differentiate between a cloned voice and a real one.

Bill Proposes AI Liability Legal Reform

Senators Josh Hawley and Richard Blumenthal proposed a bipartisan bill, the "No Section 230 Immunity for AI Act," intending to clarify that generative AI doesn't qualify for legal immunity under Section 230 of the Communications Decency Act, If enacted, the bill would permit individuals to sue companies for alleged harm by generative AI models, countering the position of some legal experts who argue that denying AI developers Section 230 protection could lead to damaging lawsuits or overly cautious product development.

More News

• Meta lost around a third of its AI researchers over the last year, causing the company to lag behind the competition in the current AI arms race. (WSJ)
• Google has proposed a regulatory framework that contrasts to those proposed by OpenAI and Microsoft. (Washington Post)
• EU's competition commissioner, Margrethe Vestager, asserts that the immediate threat of AI-induced discrimination outweighs potential extinction-level risks. (The Guardian)
• McKinsey believes generative AI can add $4.4 Trillion in value to the global economy. (NY Times)
• The EU is inching closer to passing comprehensive AI regulations. (Time)
• Snap released research findings on a new diffusion model that creates images in a few seconds. (Snap)
• Resident Advisor released a guide on the ethical way to use AI in making music. (Resident Advisor)
• Amazon is using generative AI to summarize product reviews. (CNBC)
• Over 300 people attended an experimental church service in Germany with a sermon delivered by OpenAI's ChatGPT chatbot. (Ars Technica)

Thank you for reading the Reality Defender Newsletter. If you have any questions about Reality Defender, or if you would like to see anything in future issues, please reach out to us here.