On November 20, 2024, Wyoming SBDC Network hosted CyberWyoming for an insightful 30-minute webinar, AI Products Roundup: Choosing the Right Tool for the Job. This session, presented by CyberWyoming’s Mary Keane and Brook Malone, provided practical advice on navigating the growing landscape of generative AI tools. Instead of promoting specific products, the focus was on helping businesses evaluate AI tools’ safety and effectiveness based on privacy, security, and terms of service considerations. Here’s a summary of the key takeaways.
Why AI Product Evaluation Matters
Generative AI tools, like ChatGPT, have revolutionized how businesses approach content creation, data analysis, and more. However, not all AI tools are created equal when it comes to protecting user data. This webinar highlighted the importance of understanding:
- Data Privacy Policies: How your data is collected, stored, and shared.
- Security Measures: How companies safeguard against vulnerabilities.
- Terms of Service: The implications of agreeing to a company’s terms.
Key Questions to Ask When Evaluating AI Tools
To make informed decisions, consider the following:
- Multi-Factor Authentication (MFA): Does the tool require it for account security?
- Data Usage: Is your data sold to third parties or used to train AI models?
- Encryption: Are your data transfers and storage encrypted?
- Policy Clarity: Are privacy policies transparent and frequently updated?
- Compliance: Does the product meet recognized standards like SOC, ISO/IEC 27001, GDPR, or CCPA?
- Breach Response: How does the company handle security incidents?
Case Study/Demonstration: Podcastle
Mary and Brook demonstrated how to evaluate a real AI product—Podcastle. They walked through steps to:
- Locate and analyze the privacy policy.
- Determine data collection and retention practices.
- Investigate terms of service for user rights.
- Identify third-party data sharing and security certifications (e.g., GDPR).
Resources and Best Practices
To evaluate AI tools effectively, it was recommended to:
- Use Privacy Tools: Analyze terms of service and security measures.
- Use Other Research Tools: Use resources like Futurepedia, OWASP, and Microsoft Purview.
- Watch Out For Red Flags: Watch for vague policies, lack of certifications, and opt-out consent practices.
Best practices for conducting evaluations include:
- Creating a Template: Document key privacy and security questions.
- Verification Steps: Cross-check findings with trusted sources.
- Continuous Monitoring: Keep policies updated and revisit evaluations periodically.
CyberWyoming's AI Product research Flyer
CyberWyoming researched different AI products and developed a simple framework to categorize AI tools based on privacy and security:
- Green: Privacy and security policies are robust. There is limited selling of user data.
- Yellow: Some concerns over privacy and/or security policies. For example, the policies of how user data is protected may be vague. There may be selling of user data, but users can opt-out or delete their data.
- Red: Many concerns over privacy and/or security policies. User data is sold, and users have limited or no ability to opt-out or have their user data deleted.
Get CyberWyoming’s AI Product Research Flyer
CyberWyoming’s AI Product Research Flyer is a handy resource for businesses looking to navigate the AI landscape more safely. To request the first edition of our AI Product Research Flyer or if you want us to research an AI product for your company, simply email us at info@cyberwyoming.org.
