Back to News
Models
Microsoft can't kill dogged researcher's Copilot for Word worm
iTnews·July 31, 2026·1 min read
AI Summary
Håkon Måløy's self-propagating prompt injection attack against Microsoft's Word Copilot has proven resilient, continuing to function despite months of mitigation attempts. The persistence of this exploit highlights ongoing challenges in securing AI-driven applications.
From the source
Håkon Måløy's self-propagating prompt injection attack survives months of mitigation efforts.
The full text couldn't be loaded here (the source may require a subscription).
View original at iTnewsKeep reading
ECARX-Backed Chip Maker SiEngine Secures US$200M Equity Financing to Accelerate Vertical Silicon-to-Software SDV InnovationPRNewswire · 1d agoTrump admin tries to rebuild election security infrastructure it gutted as midterms near | CNN PoliticsCNN · 1d agoThe Delhi high court ruling is unlikely to be the last word on intellectual property used for AI trainingLivemint · 1d agopyllmits 0.1.0Pypi.org · 1d ago
Was this useful?