### AI-Driven Cognitive Assistance and Surveillance

Neurodivergent users are pioneering advanced applications for general-purpose AI, effectively acting as an innovation lab for cognitive assistance. These individuals leverage tools like ChatGPT and Gemini for executive functions such as organizing scattered thoughts, managing cognitive fog, and breaking down complex tasks. AI is also used for communication support, including generating social scripts for autistic users, assisting with professional writing for those with dyslexia, and reducing the energy demands of typing. This user-driven innovation reveals a significant, untapped market for cognitive assistance tools, with the global "disability economy" commanding over $13 trillion in annual spending power.

This emergence has sparked critical debate within disability communities, centering on several key tensions:

• Assistive vs. Generative: A distinction is drawn between traditional assistive technology (AT), which facilitates a user's own creation, and generative AI, which can potentially replace the creative process.
• Authenticity and Agency: Concerns exist that AI may enforce neurotypical communication norms, undermining a user's authentic voice, and that development often proceeds without meaningful participation from disabled individuals ("Nothing About Us Without Us").
• Privacy and Bias: The use of personal data for AI personalization raises concerns about surveillance and biased systems misinterpreting neurodivergent behaviors, echoing historical discrimination.

The proposed path forward involves a convergence of traditional AT and generative AI to create systems with universal design, where adaptive support is standard. This requires AI development to follow principles of meaningful user participation, focus on removing environmental barriers, preserve user agency, and ensure robust privacy safeguards.

### AI in Government Surveillance

Separately, AI is being deployed for government surveillance with significant human rights implications. U.S. Immigration and Customs Enforcement (ICE) has awarded Palantir a $30 million contract to develop "ImmigrationOS," a surveillance platform providing near real-time tracking of individuals. The system leverages data analytics and AI to enhance immigration enforcement and deportation operations, raising concerns about the growth of authoritarian AI use and surveillance capitalism within the United States.