How can we prevent AI models from cannibalizing themselves when human-generated data runs out? Scientists say they've found the answer.
Researchers have discovered a method to prevent AI models from collapsing when human-generated data becomes scarce. By incorporating even a single human-made data point into training, AI models can maintain their accuracy and reliability. This finding addresses concerns about the potential for AI systems to produce nonsensical outputs due to reliance on synthetic data.
- ▪The study suggests that adding one human-made data point can prevent AI model collapse.
- ▪Model collapse occurs when AI systems rely solely on synthetic data, leading to inaccurate outputs.
- ▪Experts warn that high-quality human-generated data may run out soon, posing risks for AI applications.
Opening excerpt (first ~120 words) tap to expand
Technology Artificial Intelligence How can we prevent AI models from cannibalizing themselves when human-generated data runs out? Scientists say they've found the answer. Researchers have found that introducing human-made data into AI training can help to prevent AI model collapse. By Roland Moore-Colyer published 21 May 2026 in News When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works. Adding an element of human touch could be the key to avoiding AI model collapse, new research finds.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Live Science.
Discussion
0 commentsMore from Live Science
