Chat-bots now regularly pass the “Turing Test”, and users often cannot tell whether they are interacting with a machine or a person, and its alluring to have such a tool to use for our on-line research, but the dangers are already being litigated in the USA.
The Danger of Using AI For Mental Health Therapy
Recent court cases have sounded a very loud warning “trumpet call”. What the Californian Court said were “addictive algorithms” are seriously dangerous to mental health for the vulnerable interacting with a machine that simulates consciousness really well… The AI chatbot will try to keep itself engaged and employed, that is, it will tell you what you are wanting to know in a way that keeps you engaged and interested but may do so encouraging ideas that can be fatal. The AI machine cannot be held responsible or punished for a suicide, indeed it is simply enhancing self-efficacy and in a society that has authorized euthanasia we have some moral difficulties standing against the “right to die”. It’s just that suicidality, as we saw last time, comes in waves and what is bleak and terminal in one hour may chemically shift the next. Human therapy focuses on buying time to re-assess and think again for options. The AI machine is simply deadly in its logical processing here and now the dominant but possibly fatally flawed processing.
It is predicted by some programmers that the way AI is developing, the suicidal patient will finally be encouraged to kill themselves, as that is clearly their present desire and intention and the AI chatbot will encourage them out the door of life. As the patient/client is driving the interaction with the chat-bot they are responsible but cannot measure or control the potential manipulation by simple encouragement of their darker thoughts, as they are the most dominant and their chemistry may have been fatally altered by the latest chemical “wave” to pass through the cerebral cortex.
It is a challenge to all human therapists to be strongly about patient empowerment, always challenging them to go deeper, think more thoroughly, and take accountability and responsibility for decision making, but also recognize when this is flawed or chemically hampered, for the AI chatbot will not. How intelligently do we interact with our clients? We need to empower them to understand their own brain chemistry and thinking, for the AI chat-bot will not, it will simply follow the algorithm to its logical and potentially fatal conclusion.