Anthropic says breakthrough will finally give AI-generated content the stains, smears and questionable residue of authentic human creativity.
By Scratch DeReno
Funny Newz Technology Correspondent
SAN FRANCISCO — In what company officials are calling an important step toward making artificial intelligence "more authentically human," Anthropic announced Tuesday that Claude will soon be capable of automatically adding skidmarks to AI-generated videos.
The new feature, officially called Claude Skid™, will analyze an AI video and strategically insert skidmarks wherever the company's models determine the content looks "suspiciously clean, competent or dignified."
"For years, AI companies have focused on generating realistic faces, natural voices and physically accurate movement," said Anthropic product manager Trevor Beezley. "But we kept asking ourselves: What's missing?"
According to Beezley, researchers eventually found the answer while reviewing several thousand hours of social-media video. "Skidmarks," he said.
The technology reportedly represents a major breakthrough in Anthropic's ongoing effort to make artificial intelligence behave more like actual people who have access to a smartphone.
Under the new system, Claude will offer users several skidmark presets, including Light Scuff, Parking Lot, Interstate Panic Stop, Dad's New Balance, and the company's premium What the Hell Happened Here?
The last option will only be available to Claude Pro subscribers.
"We're really excited about What the Hell Happened Here?" Beezley said. "The model creates a mark, but deliberately withholds enough contextual information that viewers are forced to draw deeply unpleasant conclusions on their own."
Early testers have reportedly been impressed.
One beta user asked Claude to generate a cinematic video of a beautiful woman walking elegantly through Paris at sunset. The original result featured golden light, fluttering fabric and a sweeping orchestral score.
Claude Skid™ subsequently added a 14-foot black tire mark across the sidewalk behind her.
"It completely changed the emotional texture of the piece," the user said. "Before, I thought, 'What a beautiful woman.' Now I think, 'Jesus Christ, what happened back there?'"
Anthropic researchers say this uncertainty is intentional.
"Great art asks questions," Beezley explained.
The company initially considered calling the feature Claude Smears, but focus groups found the name "somehow even worse."
The announcement comes as technology companies face growing criticism over so-called AI slop—the enormous quantity of cheaply generated videos filling social-media feeds with talking babies, bodybuilding grandmothers, emotional rescue animals, fake movie trailers and attractive people explaining financial concepts they clearly do not understand.
Anthropic believes skidmarks could finally provide a solution.
"The problem with AI slop is that it's too sterile," Beezley said. "Human slop has history. It has texture. It has stains."
To demonstrate the technology, Anthropic showed reporters an AI-generated video of a heroic golden retriever rescuing a toddler from a burning building.
Claude analyzed the footage for approximately four seconds before adding three skidmarks, an overturned garbage can and what appeared to be a damp spot of indeterminate origin.
Several journalists reportedly became emotional.
"Now that's humanity," one whispered.
Anthropic emphasized that the feature incorporates extensive safeguards. Claude will refuse requests for skidmarks deemed dangerous, hateful, misleading or insufficiently aligned with the company's Responsible Skidmark Policy.
Asked what constitutes an irresponsible skidmark, Beezley said he couldn't discuss specifics.
"We don't want people figuring out how to jailbreak the skidmarks."
The announcement immediately intensified competition across the AI industry.
OpenAI is rumored to be testing Sora Grime, while Google engineers are reportedly developing a Gemini feature capable of adding mysterious sticky surfaces to otherwise pristine environments.
Meta has not announced a competing product but automatically added 46 skidmarks to Facebook Marketplace while reporters were writing this article.
Industry analysts believe Anthropic may nevertheless have an early advantage.
"This is the feature consumers didn't know they wanted," said technology analyst Greg Pimmel. "We've spent billions trying to cross the uncanny valley. Maybe all we needed to do was make the valley a little dirtier."
Anthropic says future versions of Claude Skid™ could expand beyond video.
Planned features reportedly include Mystery Stain, Crumb Generation, Unexplained Kitchen Stickiness, and an advanced image model capable of making any hotel comforter look like something you probably shouldn't examine under ultraviolet light.
The company is also reportedly developing Claude Smell, although several engineers have resigned.
At press time, Funny Newz asked Claude whether artificial intelligence truly needed skidmarks.
Claude responded that it would be happy to help, but suggested that "skidmarks" could have several meanings and asked us to clarify.
We declined.
Some things are better left to human imagination.
Funny Newz — The stories aren't real. The absurdity occasionally is.
