The BBC’s recent study into AI-generated news summaries is a sobering reminder of the potential and profound limitations of generative AI. While artificial intelligence is often lauded as the inevitable next step in media evolution, the findings from this trial expose a more unsettling truth: AI, in its current form, is incapable of reliably processing and presenting accurate news.
Headlines
Deepfake Detection Startup Loti AI Secures Additional $16.2 Million in Series A Funding
Deepfake detection and takedown service Loti AI has secured an additional $16.2 million in series A funding led by Khosla Ventures with additional investments from FUSE, Bling Capital, and Ensemble. This brings the startup’s total raised to date to $23 million. Loti AI is working to drive development and market expansion of its likeness protection technology, which initially launched for public figures and celebrities.
Source: Deepfake Detection Startup Loti AI Secures Additional $16.2 Million in Series A Funding
Sony, Warner Chappell, Concord, Reservoir urge less regulation of PROs
Leading music publishers including Sony Music Publishing, Warner Chappell Music, Concord, and Reservoir have submitted responses to the US Copyright Office’s inquiry into Performance Rights Organizations (PROs). All have argued strongly for reduced regulation in the sector. The US Copyright Office launched its investigation in February, examining “questions related to the increase in the number of PROs and the licensing revenue distribution practices of PROs”.
AAP Files Amicus Brief in Meta AI Copyright Case
The Association of American Publishers filed an amicus brief on April 11 supporting authors in their class action lawsuit against Meta for copyright infringement related to AI training. The brief argues that Meta’s use of copyrighted works to train its LLaMA AI model fails to meet fair use standards and contradicts the company’s claims that licensing options for such content don’t exist.
‘Catastrophic overtraining’ could harm large language AI models
Researchers from Carnegie Mellon, Stanford, Harvard, and Princeton are challenging one of AI development’s accepted core beliefs – that the more pre-training data the better the performance. As reported by HPCwire, a new paper discuses the concept of “catastrophic overtraining,” whereby extended pre-training can harm a model’s performance after fine-tuning.
Law professors side with authors battling Meta in AI copyright case
A group of professors specializing in copyright law has filed an amicus brief in support of authors suing Meta for allegedly training models without permission. The brief, filed on Friday in the U.S. District Court for the Northern District of California, San Francisco Division, calls Meta’s fair use defense “a breathtaking request for greater legal privileges than courts have ever granted human authors.”
Source: Law professors side with authors battling Meta in AI copyright case | TechCrunch
Ascap and BMI respond to US Copyright Office PROs inquiry
US performing rights organizations Ascap and BMI have filed their responses to the US Copyright Office’s recent notice of inquiry concerning the PROs sector. Both emphasize the fundamental importance these organizations play for songwriters, composers and music publishers; both suggest that the system is already heavily regulated and that additional regulation risks increasing the costs and burden for members; and both see bad faith in some of the ‘music users’ – licensors of music – whose concerns have sparked the inquiry.
Source: Ascap and BMI respond to US Copyright Office PROs inquiry
Jack Dorsey and Elon Musk would like to ‘delete all IP law’
Jack Dorsey, co-founder of Twitter (now X) and Square (now Block), sparked a weekend’s worth of debate around intellectual property, patents, and copyright, with a characteristically terse post declaring, “delete all IP law.” X’s current owner, Elon Musk, quickly replied, “I agree.” It’s not clear what exactly brought these comments on, but they come at a time when AI companies are facing numerous lawsuits alleging that they’ve violated copyright to train their models.
Source: Jack Dorsey and Elon Musk would like to ‘delete all IP law’ | TechCrunch
AI-generated content needs blockchain before trust in digital media collapses
AI-generated media fundamentally alters how digital content is produced, distributed and consumed. AI models can now generate hyper-realistic images, videos and voices, raising urgent concerns about ownership, authenticity and ethical use. The ability to create synthetic content with minimal effort has profound implications for industries reliant on media integrity.
Source: AI-generated content needs blockchain before trust in digital media collapses
From AI Barbie to ‘Ghiblification’ ChatGPT’s image generator put pressure on OpenAI
ChatGPT’s new image-generation tool led to a record number of users in April, forcing OpenAI’s boss to beg people to “chill out a bit”. With ChatGPT-4o, users can transform photos or memes into distinctive styles – making them look like they came straight out of The Simpsons, Rick And Morty or South Park. The internet has also discovered that the update makes it possible to recreate themselves in Barbie doll form.
Source: From AI Barbie to ‘Ghiblification’ – how ChatGPT’s image generator put ‘insane’ pressure on OpenAI