Fraudsters have started using AI to create fake documents claiming that artworks are genuine or legally owned, the Financial Times reports. According to art insurance brokers at Marsh, chatbots and big language models are being used to forge invoices, appraisal certificates and certificates of authenticity.
In other cases, it has not been a case of deliberate fraud, but rather AI hallucinating false references to a work of art, which the owner has taken to be true.
False documents are nothing new in the art world, but AI has made them more realistic and harder to detect.
“AI makes something that’s been going on for a long time a little easier and a little faster. You don’t have to invent a professorial expert anymore — you can just let the AI do it for you,” Harry Smith of art valuation firm Gurr Johns told the Financial Times.
To counter this trend, both insurers and appraisers are now trying to use AI themselves to review metadata and identify manipulation.
More on AI’s impact on security:
Human-in-the-loop isn’t enough: New attack turns AI safeguards into exploits
AI startups leak sensitive credentials on GitHub, exposing models and training data
No Responses