HIGH_DENSITY_DATA
FILE_REF: ai-human-compatible // VERIFIED_ENTRY
ENTRY_DATE2019_01_01
Human Compatible
01_ABSTRACT_SYNOPSIS
Human Compatible helps product leaders think beyond near-term features and into the strategic risks of advanced AI systems. Stuart Russell explains why optimizing for the wrong objective can create harmful outcomes, then connects that problem to questions of control, incentives, and system design. For PMs working on AI products, it is a useful lens for product requirements, governance, and long-term safety tradeoffs.
02_INDEX_NODES
- Why capable AI systems create product and governance risks if objectives are poorly specifiedP.042
- How control problems show up in practical product decisions, not just academic theoryP.084
- Why AI teams need stronger safety assumptions before scaling usage and automationP.126
- How PMs can frame AI requirements around uncertainty, oversight, and reversibilityP.168
PUBLICATION_DATE2019
ISBN_RECORDN/A
PAGES352_UNITS
LANGUAGEENGLISH
LEVELADVANCED
RECORDS_IDai-human-compatible
FILE_SIZE17.6_MB_RAW
Use this book as a reading hub
Branch into adjacent guides, author pages, and category pages from here.
AI & Machine Learning Products category
Browse more books in ai & machine learning products.
More by Stuart Russell
See the full reading path and related themes for Stuart Russell.
AI Strategy books
Go deeper on ai strategy through the topic page.
Machine Learning Products books
Go deeper on machine learning products through the topic page.
