BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//speak.ethtokyo.org//conference-2025//speaker//FFCTHR
BEGIN:VTIMEZONE
TZID:JST
BEGIN:STANDARD
DTSTART:20000101T000000
RRULE:FREQ=YEARLY;BYMONTH=1
TZNAME:JST
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-conference-2025-KUPG33@speak.ethtokyo.org
DTSTART;TZID=JST:20250912T113000
DTEND;TZID=JST:20250912T114500
DESCRIPTION:Cryptography can’t tell us what is “good\,” but it can pr
 ovide verifiable ground truth about AI systems. Using tools like formal ve
 rification\, zero-knowledge proofs\, and verifiable computation\, we can p
 rove claims about what model ran\, what guardrails executed\, and how safe
 ty checks were enforced. This talk explores how cryptography can’t solve
  ethics\, but can make AI behavior falsifiable\, auditable\, and accountab
 le.
DTSTAMP:20260527T011532Z
LOCATION:Plasma Stage
SUMMARY:Crypto for AI Safety: Verifying “Is”\, not deciding “Ought”
  - ventali
URL:https://speak.ethtokyo.org/conference-2025/talk/KUPG33/
END:VEVENT
END:VCALENDAR