OpenAI and AARP’s OATS bring ChatGPT workshops for older adults to 10 U.S. cities

The documented product is ChatGPT, not a newly released AI system. The OpenAI article and event pages do not specify an exact model/version used in the workshops. That matters: without a model identifier, system prompt, tool configuration, safety layer version, or account tier behavior, technical reproducibility is limited. The Denver OpenAI Academy event page says attendees received guided instruction, everyday use cases, mentor support, and resources for continued learning; it also advertised 12 months free of ChatGPT Plus for attendees. AI Skills Jam for Older Adults: Denver, CO - Event | OpenAI Academy

The workshop agenda visible for Denver, Queens, Nashville, and Montgomery County follows a similar structure: check-in and sign-in help, welcome remarks, “AI Foundations & Safety,” a break, “Practical ChatGPT Examples & Exercises,” and optional Q&A/practice time. Queens materials were bilingual English/Spanish, indicating at least some localization for language access. AI Skills Jam for Older Adults: Denver, CO - Event | OpenAI Academy

This event does not document model architecture, training data, inference configuration, system-card evaluation, or benchmark results. The technical object is therefore the intervention design: a human-facilitated, in-person AI-literacy workshop using ChatGPT as the interface.

From an implementation perspective, several design choices are documented and relevant:

1. Assisted account setup. The agenda allocates time for “Support Signing into ChatGPT,” which is important because authentication, password recovery, MFA, and app-install friction often block older-adult technology adoption before instruction begins. AI Skills Jam for Older Adults: Denver, CO - Event | OpenAI Academy
2. Hands-on practice rather than lecture-only training. OpenAI Academy describes guided instructions, everyday use cases, mentors, peer learning, and practical exercises. This aligns with a 2025 older-adult AI-literacy survey of 103 adults age 50+, which found strong interest in AI learning but challenges around understanding AI and knowing where to start, with a stated preference for hands-on learning. AI Literacy Education for Older Adults: Motivations, Challenges and Preferences
3. Safety embedded in the curriculum. OpenAI says scam detection is part of the workshop, including warning signs such as urgency, secrecy, and suspicious links. This is a defensible curriculum topic given the fraud environment, but there is no public evaluation showing that the workshop improves participants’ scam-recognition accuracy or reduces victimization. Helping older adults use AI in everyday life | OpenAI
4. Mentor-supported learning. The workshop model uses trained facilitators and mentors, which may reduce abandonment and help participants translate abstract capabilities into specific routines. This is an inference from the event design, not a measured outcome. AI Skills Jam for Older Adults: Denver, CO - Event | OpenAI Academy

Relevant research suggests that interface and instruction design matter. A 2026 arXiv study on foundation models for older-adult tech support reported that older adults’ support requests may be verbose, incomplete, over-specified, or under-specified, and evaluated prompt-chaining approaches for eliciting context, paraphrasing questions, and generating solutions. Empowering Older Adults in Digital Technology Use with Foundation Models Another 2026 study of a voice-enabled community chatbot used co-design and literacy workshops with 25 participants and reported that “glass box” education improved technical understanding and perceived transparency, while users aged 80+ still faced usability challenges. Bridging the Cognitive Gap: Co-Designing and Evaluating a Voice-Enabled Community Chatbot for Older Adults These studies do not validate OpenAI’s workshop, but they support the broader design premise: older-adult AI programs should combine usability support, transparency, and situated practice.

The workshops sit at the intersection of three trends: older adults’ rising technology use, anxiety about AI trust/privacy, and escalating elder fraud. AARP’s 2026 technology-trends research reports that interest and use of AI tools among adults 50+ are growing, while trust, privacy, and data security remain major adoption concerns. Tech Use and Adoption Growing Among Adults Age 50-Plus FBI IC3 materials report that complaints from victims over 60 exceeded 201,000 in 2025, with reported losses above $7.7 billion; this supports why scam education is a salient use case, though it does not prove ChatGPT is an effective anti-scam tool. Scammers Target Older Adult Victims — FBI

Health use is a cautionary adjacent domain. AARP reported in 2026 that 41% of adults over 50 who use AI had asked an AI-powered tool a health-related question, while many had privacy and trust concerns. Older Adults Use AI for Health Questions, But Have Trust Issues Separately, a Nature Medicine evaluation of ChatGPT Health found substantial under-triage in simulated emergency scenarios, while a later arXiv critique argued that the headline failure rate was sensitive to forced-choice evaluation format and did not necessarily represent deployed conversational behavior. ChatGPT Health performance in a structured test of triage recommendations | Nature Medicine The takeaway for this workshop is not “AI health advice works” or “does not work”; it is that older-adult AI education should explicitly teach boundaries: when to verify, when to call a professional, and when not to rely on a chatbot.

The most important limitation is lack of outcome evidence. No public materials found disclose pre/post literacy scores, scam-detection tests, longitudinal retention, participant demographics beyond “older adults,” accessibility accommodations, language coverage across all cities, or adverse-event reporting.

Second, the initiative is vendor-led. OpenAI has a legitimate interest in broadening ChatGPT adoption, and the offer of free ChatGPT Plus is both an access subsidy and a product-acquisition mechanism. That does not invalidate the program, but it makes independent evaluation important.

Third, scam-spotting use cases are double-edged. A chatbot can help explain warning signs, but users might over-trust confident outputs, paste sensitive information into chat, or treat a probabilistic response as authoritative. OpenAI’s 2025 guidance itself says ChatGPT should be combined with personal judgment and basic safety steps, not replace them. Partnering with AARP to help keep older adults safe online | OpenAI

Fourth, age-inclusive AI cannot be reduced to accessibility checklists. Research warns that older adults differ widely by health, digital experience, income, disability, language, and social support. The 2026 voice-chatbot study’s weaker usability for participants aged 80+ is a reminder that “older adults” is not a homogeneous user group. Bridging the Cognitive Gap: Co-Designing and Evaluating a Voice-Enabled Community Chatbot for Older Adults

For businesses serving older consumers, the program signals that AI onboarding may become a market differentiator. Practical training, phone registration, human mentors, accessible language, and explicit privacy/scam modules may matter as much as model capability.

For developers, the implementation lesson is to design for assisted use: clear onboarding, low-risk practice tasks, context-gathering prompts, visible uncertainty, easy escalation to humans, and privacy-preserving defaults. For enterprises and public agencies, the opportunity is to partner with trusted community institutions — libraries, senior centers, AARP/Senior Planet affiliates — rather than expecting older users to self-serve through generic documentation.

For researchers, this is a natural field-evaluation opportunity: compare workshop formats, measure retention and transfer, test scam-recognition performance before and after training, audit privacy behaviors, and publish de-identified protocols so others can reproduce the intervention.