A painting bought for less than $100 and left hanging on a wall for decades has sold for more than $250,000 after artificial intelligence helped identify its creator—offering a glimpse into how AI is changing the way people uncover hidden treasures in their own homes.
How AI Helped Uncover a Six-Figure Artwork
Read More on News
An 88-year-old woman purchased a painting for less than $100 at a New York thrift store in the 1960s. For years, it hung quietly in her home with little thought given to its origins or potential value.
That changed when her son decided to investigate the work using Google’s Gemini AI. After uploading a photograph of the painting, he asked the chatbot to identify it.
Gemini analyzed features including the painting’s color palette and Art Deco influences, and suggested it could be the work of Francis Campbell Boileau Cadell, a prominent Scottish artist of the early 20th century.
Following the AI’s suggestions, the family examined the back of the canvas for identifying marks and dates before contacting professional appraisers. Experts later confirmed the painting was an authentic Cadell.
The work eventually sold at auction for approximately £189,200, or about $254,000.
The case illustrates both the promise and limitations of AI. While technology can generate valuable leads in minutes, authentication and valuation still require expert human assessment.
It’s Not Just Paintings
The Cadell discovery is one of the most eye-catching examples, but similar stories are emerging with increasing frequency.
After purchasing a decorative porcelain plate for $4.99 at a Goodwill store in Illinois, a Chicago carpet cleaner used Google Lens to find out more about it.
Within minutes, the tool returned examples of similar pieces that had sold for thousands of dollars at auction.
Antique specialists later verified that the plate was an 18th-century Chinese export piece from the Qing dynasty, with an estimated value of between $4,000 and $6,000.
Cases like these demonstrate how AI-powered image recognition is making it easier for ordinary people to investigate items that might otherwise go unnoticed.
What once required specialist knowledge can now begin with a smartphone photo.
Think You Might Own Something Valuable? Here’s Where to Start
If you have artwork, antiques or collectibles gathering dust at home, AI tools can provide a useful first step in researching them.
Experts recommend a process like this:
Photograph the item carefully: Capture the item from multiple angles. Include close-ups of any signatures, labels, or markingsUpload to an AI tool: Use tools such as Google Lens, Gemini, ChatGPT, or Claude—basically any AI chatbot that has image capabilities. Ask specific questions, such as who made the item, when it was produced, or whether similar examples existCompare results: Check whether multiple sources identify the item in the same way. Compare with similar items and recent sale pricesSearch for provenance clues: Examine the item physically (look at the back of paintings, or on the base of objects, etc.) Look for stamps, auction labels, or signatures mentioned in AI resultsVerify the market value: Focus on completed sales rather than asking prices when researching comparable itemsConsult an expert: If the item appears significant, contact an auction house, dealer or specialist appraiser. Professional authentication remains essential before assigning any substantial value.AI Can Point the Way—But Experts Still Matter
Stories of valuable discoveries in attics, thrift stores and family homes are hardly new.
What is changing is how quickly people can investigate them.
The Cadell painting and Qing dynasty plate both show how a simple image search can uncover clues that might once have taken months or years to find.
But experts warn that AI remains a research tool rather than an authority. Authentication, provenance and valuation still depend on specialists.
As image-recognition technology becomes more sophisticated, hidden treasures may become harder to overlook—though turning a promising AI suggestion into a verified windfall will still require human expertise.