BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//bsidescharm2026//speaker//SVTPB9
BEGIN:VTIMEZONE
TZID:EST
BEGIN:STANDARD
DTSTART:20001029T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10;UNTIL=20061029T070000Z
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
END:STANDARD
BEGIN:STANDARD
DTSTART:20071104T030000
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000402T030000
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=4;UNTIL=20060402T080000Z
TZNAME:EDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
END:DAYLIGHT
BEGIN:DAYLIGHT
DTSTART:20070311T030000
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:EDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-bsidescharm2026-GBYNVR@pretalx.com
DTSTART;TZID=EST:20260426T120000
DTEND;TZID=EST:20260426T122000
DESCRIPTION:Traditional DFIR assumes that compromise produces artifacts\, f
 ailures\, or clearly malicious inputs. AI systems challenge that assumptio
 n. Models can be trained\, deployed\, and perform “as expected” while 
 still producing harmful\, biased\, or manipulated outcomes. This talk expl
 ores how data poisoning and manipulation in AI systems often target result
 s rather than content\, making traditional IOC-based detection ineffective
 . Using a DFIR mindset\, the session focuses on how investigators can iden
 tify behavioral\, temporal\, and statistical indicators that suggest somet
 hing is wrong even when no individual data point appears malicious. Attend
 ees will leave with a practical framework for thinking about AI investigat
 ions\, emphasizing baselining\, change correlation\, and forensic readines
 s over perfect attribution.
DTSTAMP:20260417T061134Z
LOCATION:Track 2
SUMMARY:Nothing Looks Broken: Investigating AI When the Model Behaves - Kia
 ra Deloatch
URL:https://pretalx.com/bsidescharm2026/talk/GBYNVR/
END:VEVENT
END:VCALENDAR