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
Within about two seconds, anyone watching this video of “Will Smith'' eating spaghetti can clearly tell that it is a deepfake, and not a particularly good one.
What is particularly interesting about this twenty second video is how much it caught fire online, as well as the new open-source model used to create it. Ars Technica has the full story behind this particularly bizarre video making the rounds.
What does Reality Defender’s platform say?
Two Calls to Pause AI Development
Top AI researchers and Elon Musk signed an open letter to call off "giant AI experiments" including (but not limited to) GPT-4. It's worth noting that Musk was one of the founders of Open AI's non-profit arm, whose for profit operation created and markets GPT4 and ChatGPT.
On the same day as the release of the open letter, the Center for Artificial Intelligence and Digital Policy alerted the Federal Trade Commission of OpenAI and GPT4’s alleged violation of the FTC’s unfair and deceptive practices.
Midjourney Halts Free Trial
After last week’s confusing Pope deepfake, along with the many deepfakes floating around created with intent to deceive, Midjourney has removed the ability to use their service for free. According to a Washington Post article on the company, the company’s leadership believe free trials are being misused, while the abusive content measures taken in recent updates have not worked.
Exposing Political Misinformation in a Leading NATO Country
Reality Defender helped a senior government official from a leading NATO country detect and disprove deepfake videos that falsely depicted them in a negative light.
Learn every step that was taken to help the country and how we can replicate it for increasingly more advanced attacks in our latest case study.
More News
- Stable Diffusion is involved in copyright lawsuits which could spell trouble for AI. (Ars Technica)
- As Bard joins the generative text race, Alphabet denies their LLM was trained on ChatGPT data. (The Information)
- AI is already being used to write fundraising emails to constituents. (The New York Times)
- Bing’s AI now has ads. (Bing)
- After chatting with an app named Chai, a Belgian man sadly took his own life. (Vice)
- Italy has banned ChatGPT. (Variety)
- A new study from Cornell shows how politicians have a hard time telling the difference between constituents and AI. (Gothamist)
- First came the Pope in Balenciaga. Now, witness what Star Wars as designed by Balenciaga (which, yes, is entirely deepfaked). (Abandoned Films)
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.
