Microsoft's msnbot suspected of AI model data harvesting
TL;DR. A website owner reports Microsoft's retired msnbot reactivated to heavily index their site after blocking bingbot from AI training. - The owner verified msnbot's genuine IP addresses and observed an indexing pattern consistent with AI data collection. - The defunct crawler reportedly began aggressive activity in 2026, despite being replaced by bingbot a decade ago. - This activity suggests unauthorized data acquisition for AI models, bypassing explicit content protection policies.
- A website owner observed renewed, aggressive indexing by the retired 'msnbot' after implementing meta tags to prevent 'bingbot' from training AI models.
- The owner verified the source IP addresses as genuine Microsoft entities, indicating a deliberate, reactivated effort.
- The indexing pattern exhibited by msnbot suggests data collection specifically for AI model training, bypassing user-defined content restrictions.
Sources
- Is msnbot being used to train AI models? — news.ycombinator.com