Uploaded January 2024 | Updated November 2026, 3 hours ago
Experiments with Midjourney and DALL-E 3 reveal a copyright minefield.
The degree to which large language models (LLMs) might “memorize” some of their training inputs has long been a question. Now, authors Gary Marcus and Reid Southen have discovered many striking instances of plagiaristic visual outputs from major generative AI platforms, plus a one-word prompt that yielded apparently infringing content: “screencap."
Read about their findings: spectrum.ieee.org/midjourney-copyright
#generativeai #ai #artificialintelligence #midjourney #midjourneyai #dalle3 #plagiarismchecker #copyright
Experiments with Midjourney and DALL-E 3 reveal a copyright minefield.
The degree to which large language models (LLMs) might “memorize” some of their training inputs has long been a question. Now, authors Gary Marcus and Reid Southen have discovered many striking instances of plagiaristic visual outputs from major generative AI platforms, plus a one-word prompt that yielded apparently infringing content: “screencap."
Read about their findings: spectrum.ieee.org/midjourney-copyright
#generativeai #ai #artificialintelligence #midjourney #midjourneyai #dalle3 #plagiarismchecker #copyright










