Player FM uygulamasıyla çevrimdışı Player FM !
Beyond the Code: Ethics and AI with Katie Paxton-Fear
Manage episode 415627677 series 3486243
Cyber Security Content Creator, Speaker & Ethical Hacker, Katie Paxton-Fear, joins Nic Fillingham on this week's episode of The BlueHat Podcast. Katie holds a PhD in defense and security AI plus cybersecurity and works as an academic, teaching undergraduate students cybersecurity topics. She also runs a popular YouTube channel focused on bug bounty hunting, hacking, and pen testing. Katie shares her journey into cybersecurity, reflects on her initial interest in undeciphered languages and how it parallels her approach to cybersecurity, both involving a fascination with solving mysteries and uncovering hidden meanings.
In This Episode You Will Learn:
- Approaching AI systems with caution when translating less-documented languages
- Concerns surrounding the use of copyrighted training data in AI systems
- Recognizing and addressing AI system limitations and biases in real-world deployments.
Some Questions We Ask:
- Can fine-tuning AI models prevent degradation and improve performance?
- What are the ethical implications of putting sensitive information into AI systems
- How does relying on niche or obscure training data impact AI models?
Resources:
View Katie Paxton-Fear on LinkedIn
View Nic Fillingham on LinkedIn
Related Microsoft Podcasts:
Discover and follow other Microsoft podcasts at microsoft.com/podcasts
Hosted on Acast. See acast.com/privacy for more information.
41 bölüm
Manage episode 415627677 series 3486243
Cyber Security Content Creator, Speaker & Ethical Hacker, Katie Paxton-Fear, joins Nic Fillingham on this week's episode of The BlueHat Podcast. Katie holds a PhD in defense and security AI plus cybersecurity and works as an academic, teaching undergraduate students cybersecurity topics. She also runs a popular YouTube channel focused on bug bounty hunting, hacking, and pen testing. Katie shares her journey into cybersecurity, reflects on her initial interest in undeciphered languages and how it parallels her approach to cybersecurity, both involving a fascination with solving mysteries and uncovering hidden meanings.
In This Episode You Will Learn:
- Approaching AI systems with caution when translating less-documented languages
- Concerns surrounding the use of copyrighted training data in AI systems
- Recognizing and addressing AI system limitations and biases in real-world deployments.
Some Questions We Ask:
- Can fine-tuning AI models prevent degradation and improve performance?
- What are the ethical implications of putting sensitive information into AI systems
- How does relying on niche or obscure training data impact AI models?
Resources:
View Katie Paxton-Fear on LinkedIn
View Nic Fillingham on LinkedIn
Related Microsoft Podcasts:
Discover and follow other Microsoft podcasts at microsoft.com/podcasts
Hosted on Acast. See acast.com/privacy for more information.
41 bölüm
Minden epizód
×Player FM'e Hoş Geldiniz!
Player FM şu anda sizin için internetteki yüksek kalitedeki podcast'leri arıyor. En iyi podcast uygulaması ve Android, iPhone ve internet üzerinde çalışıyor. Aboneliklerinizi cihazlar arasında eş zamanlamak için üye olun.