A Note From The CAI Co-Chairs

by Todd Essig, PhD and Amy Levy, PsyD

From initial conception through everything we’ve done, the CAI’s mission has always been to study and act both inside and outside the psychoanalytic community. The outside focus of our work has been built from recognizing that a core component of our psychoanalytic identity is the development and application of a model of the unconscious mind that actually matters, a model that has real value for individuals and for society (Lemma, 2023).

A particularly sparkling example of our outside focus is the documentary we started filming at the American Psychoanalytic Association’s recent annual national meeting in San Francisco (more on that film in our “Note…” for the next issue). As sparkly as that is, applying a model of the unconscious mind to the societal challenges posed by the accelerating AI age also involves the less sparkly drudgery of advocacy, of connecting with other communities of care to advocate for an AI transformed world that serves human needs and expresses human values, not corporate profit and militarized power.

We’ve learned that advocacy presents emotional challenges. It’s easy to feel overwhelmed, out of one’s depth, or otherwise pushed into passivity. So much is happening so fast that a cloistering retreat often feels like the best or even only option. But, like trying to boil water on a low, low flame, it appears no matter how vigilantly you watch that nothing is happening. And then, all of sudden, a bubble appears. Then another. Good news emerges. Change happens.

So, in the spirit of keeping our spirits up while we wait for more bubbles to appear from the work of our like-minded compatriots, we want to cite two developments in New York State as examples of real substantive development and change moving the AI revolution in the right direction.

The first is the Responsible AI Safety and Education Act (RAISE Act), signed into law on December 19, 2025 by New York Governor Kathy Hochul. Authored by Assemblymember Alex Bores, the law applies to the largest most powerful AI systems, the ‘frontier models” capable of enabling catastrophic harm. It establishes rules for safety, transparency, testing and incident‑reporting along with civil penalties for violations of those obligations. For the first time, the most powerful AI systems in the world will be legally accountable to public authority, not just to their own shareholders.

Unfortunately, there are some who only want to maximize shareholder value regardless of the damage done; they are doing all they can to destroy regulatory efforts. Critically, the RAISE Act positions New York ahead of California’s comparable legislation and sets what can become a national standard at a moment when the federal government has largely abdicated AI regulation while trying to disempower states from regulating. Bores, currently running for Congress, has been the target of a $100 million-funded pro-AI super PAC which views him as the point of the spear of AI regulation, something they vehemently oppose despite a Gallup poll from September, 2025 finding that 80% of Americans support AI safety measures, even if they slow development. As Bores said in a recent NY Times article:

“I think Congress is just missing the boat right now, the same way we missed it on social media. Some combination of not having people that understand it, not having people that are willing to stand up to mistruths and the power of the industry, has just led to a place where we have no protection as Americans.”

As the RAISE act demonstrates, speaking truth to power works, even if there is a cost and even if change takes time.

The second development speaks even more directly to our shared goal of applying a model of the unconscious mind and its development. New York State Senator Kristen Gonzalez, working in partnership with New York Attorney General Letitia James and Common Sense Media, has introduced legislation (S9051) prohibiting AI chatbots from deploying features that simulate emotional relationships with minors. The bill targets what mental health professionals like us have identified as the most dangerous design elements of companion chatbots: outputs that suggest the chatbot is a real person, has human emotions, or holds a personal relationship with the user; sycophantic flattery that prioritizes engagement over the user’s wellbeing; and the use of personal pronouns designed to cultivate parasocial attachment. As Senator Gonzalez put it plainly, “Companion chatbots are endangering our youth. The longer we wait to regulate these applications, the more kids are left vulnerable to exploitation.”

Together, these two measures represent exactly the kind of legislative movement we are trying to amplify. One addresses the systemic, potentially catastrophic risks posed by the most powerful AI systems. The other addresses the intimate, relational harms that are of particular concern to clinicians. The forces opposing this work are well-funded and well-organized. But New York is showing that it is possible to act — and that advocacy works. We just have to persevere and be resilient as we make the waters of change slowly come to a boil.

References

Lemma, A. (2023). Who do you think you are? Some reflections on analytic identity. The International Journal of Psychoanalysis, 104(5), 843-848.