ByteDance’s InfiniteYou lets users generate unlimited variations of portrait photos – The Decoder
ByteDance has developed a new approach to AI portrait generation that tackles common problems like inconsistent facial features and poor prompt following. Unlike previous solutions such as PuLID-FLUX that directly modify AI model attention, InfuseNet processes facial features as a parallel information layer. This keeps the core AI model intact while improving portrait generation quality.
Rembrandt to Picasso: Five ways to spot a fake masterpiece – BBC Culture
In authenticating the painting in the Burlington Magazine, one expert insisted “in no other picture by the great Master of Delft do we find such sentiment, such a profound understanding of the Bible story – a sentiment so nobly human expressed through the medium of the highest art”. But it was all a lie. In a remarkable twist, Van Meegeren eventually chose to expose himself as a fraudster shortly after the end of World War Two, after being charged by Dutch authorities with the crime of selling a Vermeer – therefore a national treasure – to the Nazi official Hermann Göring. To prove his innocence, if innocence it might be called, and demonstrate that he had merely sold a worthless fake of his own forging, not a real Old Master, Van Meegeren performed the extraordinary feat of whisking up a fresh masterpiece from thin air before the experts’ astonished eyes. Voilà, Vermeer.
Exclusive: These universities have the most retracted scientific articles – Nature
This surge can now be seen in a first-of-its-kind analysis of institutional retraction rates around the globe over the past decade, for which Nature’s news team used figures supplied by three private research-integrity and analytics firms. Jining First People’s Hospital tops the charts, with more than 5% of its total output from 2014 to 2024 retracted — more than 100 papers (see ‘Highest retraction rates’). That proportion is an order of magnitude higher than China’s retraction rate, and 50 times the global average. Depending on how one counts, the hospital could be the institution with the world’s highest retraction rate. Many other Chinese hospitals are retraction hotspots. But universities and institutes in China, Saudi Arabia, India, Pakistan and Ethiopia feature in the data as well. Retractions can be for honest mistakes and administrative errors, but evidence suggests the majority of cases in these data are related to misconduct. […]
Data on retractions show that they are rare events. Out of 50 million or more articles published over the past decade, for instance, a mere 40,000 or so (fewer than 0.1%) have been retracted, according to the firms’ data sets. But the rise in retraction notices (by which journals announce that a paper is being retracted) is outstripping the growth of published papers — partly because of the rise of paper mills and the growing number of sleuths who spot problems with published articles.
AI ‘inspo’ is everywhere. It’s driving your hair stylist crazy. – Archive Today: The Washington Post
When a potential client approached event planner Deanna Evans with an AI-generated vision for her upcoming wedding, Evans couldn’t believe her eyes, she said. The imaginary venue was a lush wonderland, with green satin tablecloths under sprawling floral arrangements, soft professional lighting and trees growing out of the floor. “It looked like the Met Gala,” Evans said. The idea would have run the client around $300,000, she guessed, which was four times her budget. Evans delicately explained the problem — and never heard from the woman again.
Faking It: Deepfake porn site’s link to tech companies – Bellingcat
“It’s par for the course that you’ll have a parent company and then a very long list of subsidiaries that are registered in Hong Kong, because Hong Kong has a different legal structure than mainland China,” she said. “You want six or seven levels of distance between the main parent company and then whatever company is doing the main business. This is how many Chinese companies engage in questionable behaviour.”
Deepfake videos are getting shockingly good – TechCrunch
Researchers from TikTok owner ByteDance have demoed a new AI system, OmniHuman-1, that can generate perhaps the most realistic deepfake videos to date. … According to the ByteDance researchers, OmniHuman-1 only needs a single reference image and audio, like speech or vocals, to generate a clip of an arbitrary length. The output video’s aspect ratio is adjustable, as is the subject’s “body proportion” — i.e. how much of their body is shown in the fake footage. […]
The implications are worrisome. Last year, political deepfakes spread like wildfire around the globe. On election day in Taiwan, a Chinese Communist Party-affiliated group posted AI-generated, misleading audio of a politician throwing his support behind a pro-China candidate. In Moldova, deepfake videos depicted the country’s president, Maia Sandu, resigning. And in South Africa, a deepfake of rapper Eminem supporting a South African opposition party circulated ahead of the country’s election.
AI Amazon Apple archives art book reviews books chatbots children China classical commerce computing copyright culture data design Donald Trump economics education Elon Musk environment ethics fakes future gadgets galleries Google health history history of technology humour illustration internet jobs journalism Leeds LLM mental health metaverse mobile phones movies music NASA painting photography politics productivity psychology robots satire schools science sculpture search Second Life social media society software space standards technology television time truth video video games VR watches writing
More miscellaneous details 