A few weeks ago on the radio, while driving from one place to another, I heard an interesting and scary program. The National Eating Disorders Association (NEDA) had to suspend its helpline for people suffering from eating disorders because, when they called, the chatbot powered by Artificial Intelligence was offering advice on how to go on a diet to lose weight. As soon as the radio journalists reported on it, the Association immediately shut down the program. Kate Wells tells us that NEDA exists to support patients with eating disorders. Initially, the helpline was manned by humans, but during the pandemic, the calls increased significantly, reaching 70,000 calls last year. They closed the helpline because they were overwhelmed and introduced Tessa, a robot, because patients were upset that the helpline was no longer available. Here is the original link if you want to listen to it: https://www.npr.org/sections/health-shots/2023/06/08/1180838096/an-eating-disorders-chatbot-offered-dieting-advice-raising-fears-about-ai-in-hea
There is no blame to be assigned to the situation and the reason we found it
interesting to share with you is because:
- Artificial Intelligence (AI) can only be programmed to give ‘common-sense’ responses which, when faced with relational problems, end up being ‘more of the same’. From the point of view of Problem Solving Brief Therapy those actions, with the best of intentions will not yield different/positive results. In this case, when someone calls to complain about having issues with food, the AI suggests ways to eat more healthily and lose weight because that’s what it has been programmed to do: give common sense advice.
- AI, for now, does not understand the relationships between organisms or between organisms and the environment. It has a linear perspective: if a, then b. Furthermore, it lacks the ability to listen and learn about the individual’s particular point of view with whom it is interacting. Therefore it cannot let the person know that it has understood their issue by using the language that person can hear. It lacks the possibility of ‘tailored therapy,’ which is necessary to promote positive change.
- AI in chat mode is blind to its own limitations. It cannot ‘think’ in a dimension outside the chat, as it is incapable of ‘perceiving’ information from the multidimensional world. It receives insufficient information (text-based language), and it assumes that this is the only dimension needed to provide an answer. Bateson would call this ‘Shoddy Epistemology,’ as he referred to schemes that simplified the complexity of the relationship between the organism and the environment.
We argue in this short piece that we hope that AI can never replace a human therapist, as it is not even capable of perceiving beyond its current dimension, much less can it apply the basic premises of what we teach as MRI Problem Solving Brief Therapy model. The use of Artificial Intelligence to reduce costs therefore results in even greater resource loss. With the best of intentions, the implementation of Tessa will have created further problems for the initial callers wanting to get advise. Because there already is a lack of available mental health professionals, the investment needed to treat the outcomes of AI casualties will make the situation worse. As imagined in the radio report, many of those human beings
who spoke with Tessa may have required hospital assistance, with the economic and human costs that it implies.
We hope that this way of viewing the world gives you a ‘headache’ due to its complexity! We are also hoping that you might decide ‘how interesting’ and pay us a visit in the upcoming brief training on August 11. Join us for an overview of what we are about. Of course, on the other hand if you are comfortable with your point of view, there is at least a 50% chance you will say to yourself, ‘too much hassle: I’ll stick with what I have.’ The decision, of course, is yours!
We would love to hear your thoughts or if you have alternative perspectives; those are the ones that construct ‘high-quality’ epistemologies.
