BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Europe/Istanbul
BEGIN:DAYLIGHT
DTSTART:20380119T061407
TZOFFSETFROM:+0300
TZOFFSETTO:+0300
RRULE:FREQ=YEARLY;BYDAY=3TU;BYMONTH=1
TZNAME:+03
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20160907T000000
TZOFFSETFROM:+0300
TZOFFSETTO:+0300
RRULE:FREQ=YEARLY;BYDAY=1WE;BYMONTH=9
TZNAME:+03
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260223T121104Z
UID:F0150B4F-754F-4811-A0B7-2F51B47D7A0E
DTSTART;TZID=Europe/Istanbul:20260212T140000
DTEND;TZID=Europe/Istanbul:20260212T153000
DESCRIPTION:Like any scientific field\, software security requires empirica
 l evidence to understand the capabilities and limitations of using AI tool
 s. Over the past few years\, we have conducted extensive research explorin
 g different aspects of software security for gathering\, using\, and disse
 minating empirical evidence. Our efforts aim to systematically understand 
 and address the challenges that may negatively impact the trustworthiness 
 and scalability of AI-powered solutions for engineering secure digital sys
 tems. This talk presents findings from our empirical examination of AI-gen
 erated code and software vulnerability prediction models. I will present e
 mpirical evidence showing where AI claims diverge from reality. My talk wi
 ll focus on helping audience to better understand how LLMs generated code 
 can introduce security weaknesses in software systems and what types of da
 ta quality issues result in less reliable vulnerability prediction models.
  The talk concludes with evidence-based guidance about what current AI too
 ls for code generation can and cannot do in enterprise security workflows 
 to support informed decisions about when and how to trust AI assistance.\n
 \nSpeaker(s): Ali Babar\n\nNişantepe Mah.\, Orman sokak\, Özyeğin Üniv
 ersitesi\, Istanbul\, Istanbul\, Türkiye\, 34794\, Virtual: https://event
 s.vtools.ieee.org/m/537003
LOCATION:Nişantepe Mah.\, Orman sokak\, Özyeğin Üniversitesi\, Istanbul
 \, Istanbul\, Türkiye\, 34794\, Virtual: https://events.vtools.ieee.org/m
 /537003
ORGANIZER:reyhan.aydogan@ozyegin.edu.tr
SEQUENCE:19
SUMMARY:Evidential Understanding of AI-Powered Software Security: Findings 
 on LLM-Assisted Development and Data Quality
URL;VALUE=URI:https://events.vtools.ieee.org/m/537003
X-ALT-DESC:Description: <br /><p>&nbsp\;Like any scientific field\, softwar
 e security requires empirical evidence to understand the capabilities and 
 limitations of using AI tools. Over the past few years\, we have conducted
  extensive research exploring different aspects of software security for g
 athering\, using\, and disseminating empirical evidence. Our efforts aim t
 o systematically understand and address the challenges that may negatively
  impact the trustworthiness and scalability of AI-powered solutions for en
 gineering secure digital systems. This talk presents findings from our emp
 irical examination of AI-generated code and software vulnerability predict
 ion models. I will present empirical evidence showing where AI claims dive
 rge from reality. My talk will focus on helping audience to better underst
 and how LLMs generated code can introduce security weaknesses in software 
 systems and what types of data quality issues result in less reliable vuln
 erability prediction models. The talk concludes with evidence-based guidan
 ce about what current AI tools for code generation can and cannot do in en
 terprise security workflows to support informed decisions about when and h
 ow to trust AI assistance.</p>
END:VEVENT
END:VCALENDAR