The prediction is framed on a multi‑year timeline (“over the coming years” and then trial lawyers will end access), but we are only about three and a half months past the August 15, 2025 episode date—far too early to judge its long‑run accuracy.

1. AI as a global educational leveler (especially outside the U.S.)
There is clear early evidence of AI tutors and personalized learning being deployed worldwide:

• China uses AI tutors such as Squirrel AI to provide personalized instruction and help close urban‑rural education gaps. (weforum.org)
• South Korea and China are rolling out AI‑powered textbooks and national AI curricula aimed at personalized learning for all students. (sarrauteducacion.com)
• Adoption of AI education tools is growing rapidly in many regions, and India now leads global use of Google’s Gemini AI tutoring features. (sezarroverseas.com)
  However, major reports simultaneously warn of an “AI education divide”: adoption closely tracks income levels, with very low uptake in many low‑income countries, so global educational opportunities have not yet been “significantly leveled” in a decisive way. (sezarroverseas.com) Given the explicitly future‑oriented wording (“over the coming years”), this part of the prediction cannot yet be declared right or wrong.

2. U.S. trial lawyers / regulation restricting access to AI tools via “AI psychosis” claims
The phrase “AI psychosis” (or chatbot psychosis) has indeed entered public and scientific discourse, describing delusion‑like harms tied to chatbot use. (en.wikipedia.org) There are:

• High‑profile lawsuits such as Raine v. OpenAI, alleging that ChatGPT contributed to a teenager’s suicide; media and expert commentary around this and similar cases frequently discuss “AI psychosis” and delusional reinforcement by chatbots. (en.wikipedia.org)
• Growing litigation and regulatory scrutiny of AI companions and mental‑health chatbots, including multiple U.S. state laws (e.g., Illinois’ Wellness and Oversight for Psychological Resources Act, and earlier laws in Utah and Nevada) that restrict AI from acting as a therapist, motivated partly by concerns about AI‑induced psychosis and unsafe crisis responses. (en.wikipedia.org)

However, these actions are narrowly focused on therapy / emotional‑support use cases and minors, not on educational tutoring more broadly. At the same time, U.S. educational institutions are actively expanding access to AI tools: for example, Indiana University’s 2025 rollout of ChatGPT Edu to all 120,000 students, faculty, and staff, and strong adoption of AI tutors in U.S. K‑12 and higher education markets. (worldhealth.net) There is no evidence that trial‑lawyer‑driven litigation or regulation has “ended your ability” to use AI‑based personalized education tools in the United States as of November 30, 2025.

Conclusion

• Elements of the setup Sachs describes (AI tutors spreading globally; lawsuits and laws invoking AI‑related psychological harms) are beginning to appear, but the key outcome he predicts—U.S. access to AI‑powered personalized education being shut down by trial lawyers and regulations framed around “AI psychosis”—has not occurred.
• Because the prediction explicitly concerns developments “over the coming years,” and current trends are still unfolding, it is too soon to categorize it as either right or wrong.

Therefore, the appropriate status for this prediction as of November 30, 2025 is "inconclusive" (too early to tell).