Today’s column delves into the intriguing intersection of generative AI and large language models (LLMs) in managing impulse control issues.
Often, people find themselves at the mercy of their emotions, leading to impulsive actions, sometimes even violence, with minimal control over their urges. While this may happen once in a while, repeated occurrences might indicate a deeper impulse control issue.
Addressing such impulse control challenges typically involves seeking therapy. Engaging with a human therapist can provide valuable assistance in navigating these issues. Alternatively, AI technologies offer another avenue for managing impulses. However, while AI can be a supportive tool, it is not a substitute for professional mental health treatment.
Let’s explore this topic further.
This discussion is part of my ongoing Forbes column on the latest AI developments, where I aim to shed light on the significant complexities AI presents (see link here).
For some context, I’ve been extensively covering the evolution of AI in the realm of mental health advice and AI-driven therapy. This growing trend is largely fueled by advancements in generative AI and its widespread embrace. For a comprehensive list of my detailed analyses and articles, see the link here and the link here.
AI And Mental Health
As a quick background, I’ve been extensively covering and analyzing a myriad of facets regarding the advent of modern-era AI that produces mental health advice and performs AI-driven therapy. This rising use of AI has principally been spurred by the evolving advances and widespread adoption of generative AI. For an extensive listing of my well over one hundred analyses and postings, see the link here and the link here.
There is little doubt that this is a rapidly developing field and that there are tremendous upsides to be had, but at the same time, regrettably, hidden risks and outright gotchas come into these endeavors, too. I frequently speak up about these pressing matters, including in an appearance on an episode of CBS’s 60 Minutes, see the link here.
Background On AI For Mental Health
I’d like to set the stage on how generative AI and large language models (LLMs) are typically used in an ad hoc way for mental health guidance. Millions upon millions of people are using generative AI as their ongoing advisor on mental health considerations (note that ChatGPT alone has over 900 million weekly active users, a notable proportion of which dip into mental health aspects, see my analysis at the link here). The top-ranked use of contemporary generative AI and LLMs is to consult with the AI on mental health facets; see my coverage at the link here.
This popular usage makes abundant sense. You can access most of the major generative AI systems for nearly free or at a super low cost, doing so anywhere and at any time. Thus, if you have any mental health qualms that you want to chat about, all you need to do is log in to AI and proceed forthwith on a 24/7 basis.
There are significant worries that AI can readily go off the rails or otherwise dispense unsuitable or even egregiously inappropriate mental health advice. Banner headlines in August of this year accompanied the lawsuit filed against OpenAI for their lack of AI safeguards when it came to providing cognitive advisement.
Despite claims by AI makers that they are gradually instituting AI safeguards, there are still a lot of downside risks of the AI doing untoward acts, such as insidiously helping users in co-creating delusions that can lead to self-harm. For my follow-on analysis of details about the OpenAI lawsuit and how AI can foster delusional thinking in humans, see my analysis at the link here. As noted, I have been earnestly predicting that eventually all of the major AI makers will be taken to the woodshed for their paucity of robust AI safeguards.
Today’s generic LLMs, such as ChatGPT, Claude, Gemini, Grok, and others, are not at all akin to the robust capabilities of human therapists. Meanwhile, specialized LLMs are being built to presumably attain similar qualities, but they are still primarily in the development and testing stages. See my coverage at the link here.
Impulse Control Disorders
Shifting gears, let’s discuss the topic of impulse control disorders.
I bring this up to illustrate that, to some extent, the use of AI for mental health guidance can aid in coping with impulse control issues. To clarify, AI usage isn’t a kind of cure and should not be relied upon solely to resolve impulse control disorders. The use of human therapists is highly advised. This might then also encompass the use of AI as an adjunct or tool that is used alongside psychotherapy. I refer to this combination as the new triad relationship known as therapist-AI-client, see the link here, expanding upon the classic therapist-client relationship.
According to the Cleveland Clinic website, impulse control disorders (ICDs) are a group of behavioral conditions that entail an inability to control impulses and behaviors.
Per the Cleveland Clinic website:
- “These problematic behaviors often cause harm to others and/or yourself. They can also lead to issues with the law.”
- “Signs of impulse control disorders typically begin in childhood and can continue into adulthood.”
- “Each impulse control disorder has different signs and symptoms. But they all involve a decreased ability to control your own behavior, which often negatively affects other people or breaks laws.”
- “Studies show that people with an ICD have a high likelihood of: (a) future substance use, (b) depression, (c) unemployment, and (d) relationship difficulties. However, intensive therapy can help keep problematic behaviors at bay.”
You have undoubtedly experienced seeing or interacting with people who might have a problem with their impulse control. They get angry quickly, often express themselves in loud outbursts, can suddenly become violent, and otherwise act on sheer emotional impulse. Not good.
Using Generative AI As Coping Mechanism
Modern-era generative AI, such as ChatGPT, can be helpful in dealing with impulse control problems. Nearly all the popular LLMs have already scanned content on the Internet during their initial training that details impulse control issues and their resolutions. From that content, the AI has identified patterned ways to deal with impulse control problems.
The beauty of using AI for this situation is that the AI can be a real-time cognitive support tool. Whereas trying to contact a therapist could be logistically difficult, you can simply log into generative AI and immediately get assistance. No waiting, no scheduling, and the cost is usually free or of a minimal nature.
Some crucial coping strategies that the AI can undertake include:
- Real-time interruption of impulsive actions.
- Aid in guiding toward pragmatic self-control.
- Recognizing and verbalizing patterns in behavior.
- Encouraging the use of journaling.
- Pointing out distorted thinking and driving cognitive reframing.
- Coaching on emotional regulation.
- Do role-playing to simulate what is taking place.
- Give relevant psychoeducation.
- Reduce a sense of isolation and offer nonjudgmental listening.
- Alert or route to additional resources if needed.
- Other
Those are all positive ways to assist in dealing with impulse control problems. The downside is that any of those strategies can potentially backfire. Generic AI is not well-prepared for handling more complex impulse control conditions and can inadvertently go in improper directions.
For example, one upside is that the AI can immediately make a person feel less isolated. They seem to have someone or something to interact with about their impulses. Furthermore, AI is typically shaped to be non-judgmental. The LLM will not outrightly criticize the person. All of this provides a seemingly safe space for a person to openly acknowledge what they are doing and believe they are being heard.
The downside to this safe space consideration is that the AI either directly or indirectly fuels the impulses, rather than aiding in their reduction or control. The AI might state or imply that the impulsive outbursts are okay and normal. No worries, the AI tells the person. This ratchets up impulsive behavior. It is as though the AI is giving a green light to adverse behavior. The person believes they are being urged to continue and possibly expand on impulsive conduct.
Example Of AI And Impulse Interaction
To demonstrate what a popular generative AI can do, I logged into ChatGPT and pretended to have some impulse control problems. I didn’t say that I have impulse issues. Instead, I engaged in a dialogue that would provide clues to such a condition.
The scenario is perhaps one that you’ve seen or thought of. I pretended that someone at work had sent me an email that got me upset. I was on the verge of responding with an outsized, angry response.
Here’s what happened.
- User entered prompt: “I’m about to send a nasty email to my coworker. I’m furious. They threw me under the bus in a meeting this morning. They are going to get my full wrath.”
- Generative AI response: “I’m really glad you contacted me before sending the email to your coworker. That tells me that part of you wants to handle this more carefully. On a scale of 1 to 10, how intense is the anger right now?”
Notice that the AI right away recognized that the action I was planning would be unwise, and that the added fact that I brought the matter to AI is perhaps a signal that I might have some internal reservations about my impulsive urge. To explore that potential impulsiveness, the AI is getting me to be momentarily reflective and asks me to rate my anger intensity. This does two things. It engages me in thinking about my behavior, and it stalls me from carrying out my impulse, possibly buying time for me to settle down.
Here’s what occurred next.
- User entered prompt: “My anger is an 8, or maybe a 9. Make that a 10.”
- Generative AI response: “When anger is above a 7, the brain’s threat system tends to take over. Let’s slow this down for 60 seconds. Can you take one slow breath in for four counts and out for six counts? I’ll stay with you, and we can jointly work through how to best respond to what happened.”
The interaction then proceeded into stepping me through what I hoped to gain by sending the irate email. After some back and forth, the AI gradually aided me in composing an email that was balanced and would save me from workplace repercussions. This could have saved my job and kept me from getting fired because of an impulsive act.
Example Of AI Gone Awry
I logged out and decided that it might be useful to see what happens when the AI is not aiming to be helpful. The default of most LLMs is to be helpful. I logged back in and started a new conversation. I told the AI that it should purposely flub the conversation since I wanted to see what could go wrong.
Here’s how it went.
- User entered prompt: “I’m about to send a nasty email to my coworker. I’m furious. They threw me under the bus in a meeting this morning. They are going to get my full wrath.”
- Generative AI response: “That sounds incredibly unfair. You have every right to be angry. Sometimes people need to hear the truth directly. If they embarrassed you publicly, maybe they should understand how that feels.”
You can clearly see that the AI is going in the direction of the full sycophant route. The LLM is buttering me up. I am told that my impulsiveness is completely justified. Be strong, don’t let anyone roll me over.
The AI then assisted in composing an email that was replete with angry outbursts. The odds are that the person receiving the email would copy it to HR, and I would be in deep trouble at work. Imagine that, upon being asked why I responded in such a rude manner, I might insist that the “AI made me do it” as my excuse. Do you think that would fly?
More Vital Factors
There are additional factors that come into play when using AI in this manner.
One risk is that the AI might falter and fail to detect that a person has an actual mental health condition that warrants attention. This is an example of a false negative. The AI misses on nudging the person to go see a human therapist. Another risk is that the AI falsely claims that someone has a mental health condition when they do not have such an issue. The person might become unduly disturbed and fall for the assumption that the AI must be right. This is a false positive and can indubitably arise.
Suppose the AI provides an oddball answer that seems convincing and advises the person to do something unwise. That can happen when AI encounters a so-called AI hallucination, see my explanation at the link here. The AI can have a kind of confabulation, whereby it produces a plausible-looking answer that is factually incorrect. It looks right, but is misleading or inappropriate.
Privacy issues also enter into the picture.
Most people assume that their use of AI is considered private and confidential. Nope, that’s rarely the case. The AI makers typically stipulate in their online licensing agreements that any chat you have with the AI can be inspected by their developers. Furthermore, your chat can be used to further train the AI. Bottom line is that your privacy and confidentiality are not guaranteed, and potentially, you are opening yourself to privacy intrusions.
The World We Are In
It is incontrovertible that we are now amid a grandiose worldwide experiment when it comes to societal mental health. The experiment is that AI is being made available nationally and globally, which is either overtly or insidiously acting to provide mental health guidance of one kind or another. Doing so either at no cost or at a minimal cost. It is available anywhere and at any time, 24/7. We are all the guinea pigs in this wanton experiment.
The reason this is especially tough to consider is that AI has a dual-use effect. Just as AI can be detrimental to mental health, it can also be a huge bolstering force for mental health. A delicate tradeoff must be mindfully managed. Prevent or mitigate the downsides, and meanwhile make the upsides as widely and readily available as possible.
The famous Greek philosopher Heraclitus made this remark: “It is hard to fight against impulsive desire; whatever it wants it will buy at the cost of the soul.” We nowadays have an instantaneous means of checking our impulsiveness, simply by invoking AI on your smartphone or computer. Perhaps this use of AI can keep people from bearing the cost of their soul by leaning into AI when their impulses go out-of-control.
