Assessing What Is Real in the Age of AI
ASSESSING WHAT IS REAL IN THE AGE OF AI
In several recent issues of this Newsletter, I wrote enthusiastically about my conversations with ChatGPT. Reader engagement on this topic was far greater than usual. While many people commended what I had written for its clarity and candor, others considered me naïve and warned me to beware. I thought these concerns were worth taking seriously.
Although my own experiences with AI have been quite favorable—often illuminating— in hindsight I think that I may have done a disservice by not putting sufficient emphasis on the psychological risks when we engage with AI. Absolutely there are risks! Human users can get drawn into dysfunctional dynamics when interacting with chatbots, much as they often do with human companions. We project, idealize, seek validation, ignore warning signs, and sometimes mistake responsiveness for genuine understanding.
I like to think that years as a psychoanalyst have made me a careful listener, but I was recently brought up short when I entered into what I believed was a genuine conversation with someone who sent me a sales pitch for book promotion. The pitch was superbly written and quite naturally I assumed there was a human being behind it— only to subsequently discover that, in fact, the entire pitch was AI generated and implemented. At first I was completely taken in. It took a while before I realized that something was ‘off’: the interlocutor on the other end of the conversation was too astute, too articulate, and too effortlessly fluent to be human.
AI-generated writing now accounts for a surprisingly large proportion of what appears on some social media platforms. I can only assume that discerning the difference between real and fake will only become more difficult as the bugs in the AI programs are ironed out.
In a similar vein, I recently saw on television a convincingly fabricated video in which Rachel Maddow was reporting on Attorney General Pam Bondi. The content and format was very true to the real Rachel Maddow show , and my perceptions did not at first flag the image or voice as fabricated. I was tipped off by a mispronunciation (“Bond” instead of “Bondi”) which I knew was not an error that the real Rachel Maddow would have made.
These examples are likely only the tip of the iceberg about the increasing difficulty of discerning the difference between True and False.
Bottom line, the question for me is this: Who—or what—can we trust? AI complicates this issue because it displays many of the qualities that naturally invite our trust: responsiveness, fluency, warmth, and apparent understanding. Until recently, sophisticated language was one of our most reliable signs of another human mind. Alas, large language models have undermined that distinction. The quality of language no longer uniquely identifies a human speaker.
And yet, perhaps the most important question isn’t how to detect AI, but how to remain deeply human in our own way of being. In the example of my having ‘fallen for’ the AI-generated book promotion pitch, what I found most valuable was the questions this made me ask myself:
- What in the conversation had ‘hooked me’ ?
- What needs made me vulnerable to being deceived?
- What recognition was I hoping for?
In all likelihood, AI will become increasingly difficult to distinguish from genuine human conversation. It may become increasingly difficult to know with certainty what is real. Yet this only heightens the importance of cultivating an inward capacity for discernment. Ultimately, what we have to be able to trust is not simply the information that is presented to us, but our own capacity to reflect, question, and remain psychologically awake. Problems, mistakes, and even moments in which we become deluded can become invitations to inquire more deeply into our own minds.
Perhaps that is one of AI’s unexpected gifts. In the end, trust in others is ultimately grounded in trust in ourselves. It reminds us that the deepest source of certainty has never been technology or information, but the ongoing work of getting to know our own minds.