top of page
Search

New LGBT Tech Roadmap: Civil Rights Governance for Artificial Intelligence

Artificial intelligence is increasingly shaping how people access information, services, opportunity, and community. AI systems are being used to screen job applicants, assess tenants, support healthcare decisions, moderate online content, verify identity, generate synthetic media, and power chatbots that people may turn to for information or support.


AI systems are being deployed in a world where discrimination, surveillance, harassment, data misuse, and unequal access already exist for members of the LGBTQ+ community. Without strong governance, automated systems can reproduce those harms at scale, make them harder to detect, and leave affected people with few meaningful ways to understand or challenge the outcome.




This report offers a public policy framework for federal and state policymakers working to ensure that AI systems are accountable, transparent, privacy-protective, and consistent with civil rights values. It is not draft legislative text. Instead, it is a roadmap for how lawmakers should evaluate AI systems based on their function, context, affected communities, and real-world consequences.




The report identifies core principles for rights-protective AI governance. It calls on policymakers to, in part:

  • regulate AI based on function, risk, and consequence;

  • affirm that existing civil rights laws apply when AI systems shape or influence decisions;

  • preserve stronger state, local, and sector-specific protections;

  • require lifecycle accountability across development, deployment, monitoring, correction, and withdrawal;

  • ensure notice, explanation, human review, appeal, and redress in high-risk settings;

  • protect sensitive data, including sexual orientation, gender identity, biometric data, health-related data, precise location, youth data, and inferred identity data;

  • prohibit harmful identity inference, automated gender recognition, and biometric surveillance in high-risk contexts;

  • ensure AI systems are tested for disparate impact, proxy discrimination, accessibility barriers, and differential performance across affected populations;

  • protect access to lawful LGBTQ+ information, health resources, crisis support, and online community;

  • reject safety approaches that rely on expanded surveillance, forced identity disclosure, real-name systems, or default parental access to sensitive identity-related activity.


The report also identifies high-risk sectors where stronger safeguards are needed, including consequential decisions, privacy and training data, biometrics and automated gender recognition, content moderation and search, generative AI abuse, chatbots and mental health-adjacent tools, and law enforcement and surveillance.



LGBTQ+ communities must be part of these conversations from the beginning. Too often, the people most exposed to technology-related harms are excluded from the design, deployment, and governance of the systems that affect them. A civil rights approach to AI requires meaningful consultation with affected communities, including LGBTQ+ people, transgender and nonbinary people, LGBTQ+ youth, people with disabilities, communities of color, immigrants, low-income communities, and others facing overlapping forms of discrimination or surveillance.


LGBT Tech’s new report, Civil Rights Governance for Artificial Intelligence, provides policymakers, advocates, companies, and civil society partners with a practical framework for evaluating AI systems and building stronger protections for LGBTQ+ communities and the broader public.

bottom of page