← Back to People
Researcher at DeepMind
4
Mentions
4
Articles
4
Outlets
May 2026 — Oct 2026
Coverage
Neutral
Sentiment
Companies Covered
Appears In
Sentiment Distribution
Based on 4 scored articles
Positive 25%Neutral 25%Mixed 25%Negative 25%
Articles
Most recent firstMentioned In
Referenced in coverage
AI #191: The Hackening Begins
Don't Worry About the VasemixedOct 22, 2026
“Krakovna noted that perpetual non-disparagement clauses were not in contracts when she joined in 2016 but were introduced around the DeepMind-Brain merger in early 2023.”
Cheatsheet: looking for inexplicable AI reward hacking
Arb Research NewsnegativeOct 10, 2026
“Vika Krakovna created the important resource Specification Gaming Examples in AI, tracking instances of AI reward hacking and misalignment.”
Why Alignment Evals Need Calibration
LessWrongcautiousJul 8, 2026
“Published a list of specification gaming examples at DeepMind containing hundreds of cases where AI systems satisfy objectives while violating designer intent.”
DeepMind on automation of alignment research
SubstackcautiousMay 4, 2026
“Co-author of DeepMind paper on automation of alignment research”