Deepfake of the Week
With the rise in AI-generated and deepfaked music in the last few days alone, the previous week’s news on the AI-generated Drake/Weeknd song seems like a blur. This week saw the release of AISIS, an album of AI-generated music that sounds eerily similar to the now-defunct band Oasis, formerly led by brothers Liam and Noel Gallagher. With a deepfaked voice sounding like the former brother, this 37-minute album simulates what the band would sound like if they never had an acrimonious split. As detailed by the creator, “We’re bored of waiting for Oasis to reform, so we’ve got an AI-modeled Liam Gallagher to step in and help out on some tunes.”
What does Reality Defender’s platform say?
What does Liam Gallagher say?
Grimes Also Welcomes Her Deepfake Voice Double
Claire Boucher, also known as the musician/artist Grimes, has invited people to use her AI versions of her voice in music and “split [any] royalties 50%.” No stranger to AI-related developments, the artist stated “I think it's cool to be fused w a machine and I like the idea of open sourcing all art and killing copyright.” She later clarified: “we may do copyright takedowns ONLY for rly rly toxic lyrics w grimes voice."
Stability AI Launches LLM
Stability AI introduced a new family of open source AI language models called StableLM, aiming to offer an alternative to ChatGPT. The models, available in various parameter sizes, are released under the Creative Commons BY-SA-4.0 license and are designed to be transparent, accessible, and efficient while delivering high performance. As with the previous week’s release of the open-source Dolly, critics of open-source LLMs fear users with malicious intent can and will use them to create hateful and damaging content that ChatGPT, Bing AI, and other closed-source models are prevented from generating due to built-in safeguards.
AI Job Loss for Thee, Not For Me
Sixty-two percent of Americans believe AI will significantly affect the American workforce as a whole, yet only 28 percent anticipate being personally impacted. This is according to a new Pew Research Center survey and report, filled with a wealth of (speculative) data on what the future of AI could look like.
More News:
- Director Joe Russo of Marvel and Community fame expects to see the release of a movie fully made with generative AI and deepfakes within two years. (A/V Club)
- Snapchat released their own AI-powered chatbot, and it’s already causing issues. (Vice)
- Stack Overflow is charging companies with LLMs for usage of its data during training. (Wired)
- Reddit is also planning to charge for usage of its API in training LLMs. (The New York Times)
- Twitch and OnlyFans star Indiefoxx is the latest notable name targeted with unauthorized deepfake pornography. (The Sun)
- A researcher working on Elon Musk’s TruthGPT was arrested in an alleged domestic violence incident. (Business Insider)
- Moderna and IBN are using generative AI to advance mRNA research and development. (Fast Company)
- Microsoft and Epic are collaborating on applying the Azure Open AI Service to the healthcare industry. (Microsoft)
- Noted music industry insider Bob Lefsetz has chimed in on AI-generated music. (The Lefsetz Letter)
- Google will use generative AI to help create ads. (Ars Technica)
- The Washington Post discovered the “secret list” of sites used to train LLMs like ChatGPT. (The Washington Post)
- Microsoft allegedly has an AI Chip in the works. (The Information)
- Former Formula One driver Michael Schumacher is suing a German tabloid for an allegedly deepfaked interview. (Engadget)
- The photographer who won a photo competition with a generated image declined his award and highlighted the positives of AI in art. (The Guardian)