Artificial Intelligence Security

(AI) security challenges are usually virtualization, cloud, big data, application, database, structure and unstructured data security challenges all rolled into one big complex project.  Unauthorized changes to your neural network weights or AI program code can happen.  AI systems can be DDoS and suffer from common big data service threats that need modern controls to protect them.
When designing AI systems, you still need to perform basic security best practices and understand security within the OSI model. Your focus should be understanding the business and regulatory needs; thus, the business security requirements

However, the following should help as a guide in your AI security considerations:

Issues with AI is Big Data Collection
  • Double Digital Footprint
  • Data Collection & Storage
  • Sensitive Data Collection
  • Application Data Entry Errors
Initial Security Requirements
  • Reginal Legal & Regulatory
  • Access Controls (MFA, privilege vs. users)
  • Activity Monitoring & Alerting (logging)
  • DR & COOP
  • Mix Platforms (VMware, Docker, etc.)
People Challenge
  • Communication
  • Lack of specialized skills
  • God complex (SSH keys, access to everything)
  • Security as an after thought
Data Challenge
  • At Rest
  • In use
  • In transit
  • Physical
Environment Challenge
  • Hybrid Cloud
  • Mix solutions for patch work security
  • PROD & Non-PROD mixing
Business OPS Challenge
  • Business requirements
  • Interview stakeholders
  • Project Management
  • Security Lifecycle Development
  • Security Strategy
Where to Start?
  • Talk more
  • Interview business units
  • Design a strategy
  • Implement Security Architecture
  • LinkedIn
  • Twitter
Copyrights  2020 by CyberSEC Geek, Inc. All Rights reserved