While AI offers remarkable advancements and transformative potential, it also brings forth a range of concerns, particularly regarding biases. The potential for both representative harm and allocative harm loom as pressing issues that demand attention.
MassTLC chatted with Shanae Chapman, Founder, CEO, & Managing Director of Nerdy Diva, for a rundown on the uses and risks of AI tools & how companies can create more inclusive practices.
What type of tasks are AI tools like ChatGPT, ChatSonic, and Amazon CodeWhisperer good for?
- Coding algorithms for apps without security vulnerabilities (review required)
- Copywriting for a general audience without security vulnerabilities (review required)
- Summarizing general content without security vulnerabilities (review required)
- Brainstorming numerous ideas without security vulnerabilities (review required)
- Checking keywords and SEO for websites and apps
- Writing a draft for your resume or cover letter
- See how other people are using ChatGPT here
What are risks to using AI tools like ChatGPT?
- Biases in AI
- Allocative harm in natural language processing
- Representative in natural language processing
- Security Risks
- Sensitive company data is collected and shared throughout the database and is not encrypted.
- Vulnerabilities through integrations, APIs, sharing sensitive data across unsecured networks
- Lack of security training for using AI tools
- Biases in AI
Can companies use AI as pseudo users for testing products instead of doing research with human users?
Qualitative data can not be substituted with AI and needs to be completed with actual target users based on criteria established by PMs, UX, Engineering.
How can companies best share research data and design recommendations internally for influencing data driven decisions?
It’s important to know your audience. Some organizational cultures are more open to lengthy reports and others would prefer something closer to an executive summary. I started getting my reports to no more than 5 pages and provided links throughout to details in a separate document for easy scanning by sharing the highlights and recommendations clearly.
What can companies do to create more inclusive practices with AI and other technologies?
Create inclusive guidelines for natural language processing, style guides, Dev Portals, documentation, and design systems.
- Follow the Inclusive Tech Glossary from Nerdy Diva on GitHub
- DAIR – Dr Timnit Gebru https://www.dair-institute.org/
- Dr Joy Buolamwini https://www.ajl.org/
Lead inclusive research to assess and address biases in AI and machine learning
- Safiya Noble Algorithms of Oppression: How Search Engines Reinforce Racism
- MassTLC Juneteenth Resources https://www.masstlc.org/juneteenth-resources/
Hire and retain diverse talent to contribute to inclusive governance and standards
- BlackComputeHer https://blackcomputeher.org/
- Black Women in Data https://www.blackwomenindata.com/
- NSBE https://www.nsbe.org/
- ITSMF https://itsmfleaders.org/
MassTLC member, Nerdy Diva, is a boutique, Black woman-owned business offering world-class research, design, tech training, and content strategy for growing and scaling into mature businesses with excellent user experience, strategic diversity and inclusion, and investment in growing their community locally and globally.
Read more from Shanae in the Nerdy Diva Newsletter Archives