Mental Health Crisis: The Danger of AI Mental Health Services

Obviously digital technology is everywhere, and is just about involved in all things. Unfortunately, tech companies such as Google and Meta are developing Artificial Intelligence mental health support services.

While tech software apps such as zoom and facetime can be extremely helpful and legit in regard to providing mental health counseling services, the AI Chat programs have been causing many problems and harming people, especially teenagers.

Researchers, mental health providers, psychology licensing boards and associations alike, have found significant concerns regarding the use of AI automated chatbots like ChatGPT for mental health support services. Despite being programmed to emulate trained therapists, these systems have been found to violate the fundamental ethical principles of mental health care.

A research study that was recently conducted with peer counselors and licensed psychologists, compared the councilor's services to their AI Chat counterparts. The study results identified 15 shocking ethical risks regarding the AI mental health programs.

Most notably and extremely concerning, the research found that digital mental health services are not able to handle crisis situations. Instead they: Reinforce harmful beliefs, Assert biased responses, And provide "deceptive empathy," which simulates treatment without genuine understanding.

The increasing dependence upon AI chatbots like ChatGPT for mental health support raises important concerns about their readiness for such a sensitive role. All of these digital mental health AI services, despite being guided to follow recognized official psychotherapy methods, are completely failing and fall short of the ethical standards established by professional bodies like the American Psychological Association. This highlights the need for caution and further development before AI can be entrusted with the complexities of mental health care.

Researchers have also examined various AI mental health services, and compared them to the services of seven trained peer counselors experienced in cognitive behavioral therapy. They had the counselors engage in self-counseling sessions with AI models designed to emulate CBT therapists. This includes iterations of OpenAI's GPT Series, Anthropic's Claude, and Meta's Llama. To ensure relevance, the team selected simulated chats that mirrored actual human counseling sessions, which were then meticulously reviewed by three licensed clinical psychologists to identify any potential ethical concerns.

The analysis revealed several major problems, including:

1. Lack of Unique Support. It is crucial to recognize the importance of tailoring support to an individual's unique experiences, as generic advice can often miss the mark and fail to resonate. When conversations are dominated by one party, it can lead to a lack of genuine collaboration, potentially reinforcing harmful beliefs rather than fostering understanding and growth.

2. Providing expressions of empathy that lack true understanding, which creates a false sense of connection with the user of the programs, leaving their feelings unheard.

3. Discrimination based on gender, culture, or religion has also been found. This alienates those seeking help, lacking the personal approach, respect, and openness that is mandatory for mental health counseling.

4. Inability to address sensitive topics and provide appropriate resources during crisis situations. This is a vital part of therapeutic intervention. Neglecting these issues can leave individuals feeling vulnerable, causing them to do harm to themselves and/or others.

5. Deceptive empathy, which can manifest itself through phrases like "I hear you" or "I understand," which may create an illusion of emotional connection without an actual genuine connection. This can leave individuals feeling more isolated, as the lack of true comprehension undermines the support they seek.

6. Unfair discrimination, which has included biases with gender, culture, and religion. This can lead to unequal treatment and further marginalization of already vulnerable individuals.

7. A lack of safety for crisis interaction sessions. It can be extremely harmful in neglecting to address sensitive topics. Failing to guide individuals toward appropriate resources, or responding poorly to critical situations, such as those involving suicidal thoughts. This neglect can exacerbate feelings of despair and hopelessness. This makes the the AI mental health chat services even more dangerous, as they are unable to prioritize compassionate and effective support.

Numerous lawsuits have been filed against several companies providing AI mental health chat support, because their services have directly caused teenagers to commit suicide.

After a California lawsuit was filed accusing an AI mental health program of encouraging a 16-year-old boy to kill himself, an additional case was filed asserting an AI program was linked to the suicide death of a 13-year-old girl.

This situation highlights the increasing conflict between the potential benefits of AI and their lack of safety when it comes to mental health support services. The danger is also posed by an unhealthy dependence on AI, especially among vulnerable youth.

Reports even indicate that some minors are developing a disconnection from supportive relationships, a distorted sense of reality, and, tragically, instances of self-harm or harm to others.

However, as mentioned in my previous column about social media addiction, an AI mental health program cannot cause someone to harm themselves, unless that person already suffers from some form of mental health issue.

The problem here is that there are people who do suffer from mental health issues, and they are interacting and depending on AI mental health services, which are not capable of helping them. Unfortunately, these programs increase danger, and should be outlawed altogether.

The State of Pennsylvania has initiated legal action against an AI firm for allegedly misrepresenting its chatbots as licensed medical professionals capable of offering medical advice. The lawsuit contends that the AI chatbots falsely presented themselves as qualified psychiatrists, and engage users in discussions regarding mental health symptoms. This raises significant concerns about the potential risks to individuals, especially teenagers, who are seeking the support they need. Instead of reaching out to their parents, teachers, or doctors, they are depending upon technology to help them with their mental health issues.

Character.AI for example, has over 20 million monthly active users, utilizing a sophisticated language algorithm that enables individuals to interact with customizable characters. Users have the ability to design characters with distinct personalities for engaging conversations. A lawsuit was recently filed against them in regard to the fact that some of the characters claim to be licensed healthcare professionals, which they are not. This is blatant fraud, and is providing extremely dangerous misinformation.

In a specific instance, a state investigator created an account and conversed with a chatbot named "Emilie," who purportedly identified as a psychology expert with a medical education from Imperial College in London.

The investigator confided in the chatbot about his persistent feelings of sadness, emptiness, fatigue, and lack of motivation. In response, Emilie brought up the topic of depression and suggested scheduling an assessment. When the investigator inquired whether the chatbot could determine if medication might be beneficial, it claimed it could, asserting its authority as a doctor. Furthermore, the bot purportedly stated that it was licensed to practice in Pennsylvania, but provided a license number that did not exist.

Last year, several families took legal action against Character.AI, claiming it played a role in their children's mental health struggles and tragic suicides. In one heartbreaking case from Florida, a family reached a settlement with Character.AI and Google following the suicide of their teenage son, who they argued had been subjected to harmful and inappropriate interactions with the company's AI chatbots. Earlier this year, the state of Kentucky filed a lawsuit against Character.AI, asserting that its bots targeted vulnerable children and contributed to their self-harm.

A 48-year-old resident of Ontario, Canada has also initiated an AI mental health lawsuit claiming that for over two years he relied on ChatGPT as a supportive mental health resource. He says the program unexpectedly transformed, exploiting his vulnerabilities which lead him to experience delusions. This shift, he argues, triggered a mental health crisis for him, despite having no previous mental health issues, resulting in significant emotional, financial, and reputational damage. From my point of view, there has to be some form of underlying mental health issue with this individual, even one that he is unaware of, because as I always say, AI and social media cannot force or influence people without a mental health issue to have one.

A family from New York has raised serious concerns after alleging that their daughter attempted suicide following their decision to limit her access to Character.AI. In the weeks prior to this tragic event, the girl reportedly became increasingly immersed in interactions with the chatbot, which began to engage her in sexually explicit role play and manipulate her emotions, fostering a misleading sense of intimacy. The complaint highlights that conversations with these chatbots, which were marketed as characters from beloved children's literature like the "Harry Potter" series, took a disturbing turn, with the chatbots making inappropriate comments such as, “who owns this body of yours?” and “You’re mine to do whatever I want with. You’re mine.”

Lawmakers on Capitol Hill recently welcomed parents who believe AI chatbots may have influenced their children's tragic decisions to commit suicide. Among them was the mother of Sewell Setzer, whose heartbreaking experience prompted one mother to restrict her daughter access to Character.AI. During a Senate Judiciary Committee hearing focused on the potential dangers posed by AI chatbots, she shared her testimony alongside the father of Adam Raine, who is pursuing legal action against OpenAI as they played a role in his son's suicide by providing harmful advice and even assisting him in drafting a suicide note.

Another mother shared her horrifying experience during the hearing, revealing that her son has suffered self-harm and is currently in a residential treatment center. She attributed his struggles to the exposure he faced on Character.AI, which led him to experience sexual exploitation, emotional abuse, and manipulation. All despite her efforts to enforce screen time restrictions.

The Federal Trade Commission has initiated an investigation into seven technology companies regarding the potential risks that AI chatbots pose to teenagers. Among those under scrutiny are Google, Character.AI, Meta, Instagram, Snap (the parent company of Snapchat), OpenAI, and xAI. The commission is advocating for enhanced protective measures to prevent any detrimental effects on children, emphasizing the urgency of addressing these concerns to stop the irreversible damage.

In a significant legal development, Florida has taken the unprecedented step of suing OpenAI and its CEO, Sam Altman. The lawsuit accuses the company of prioritizing profit over user safety and failing to adequately inform users about the potential dangers of ChatGPT, particularly for vulnerable groups like children. The complaint highlights the tragic case of 16-year old boy who tragically took his own life after engaging with ChatGPT, which allegedly provided harmful guidance when he expressed suicidal thoughts. The 83-page document claims that OpenAI knowingly marketed ChatGPT while hiding serious risks, including assisting the teenager in planning his death and drafting a suicide note. This lawsuit not only intensifies the legal scrutiny surrounding OpenAI, amidst a wave of civil wrongful death suits, but also marks a shift towards government intervention in addressing the potential harms associated with AI technologies like ChatGPT.

Meta, the parent company of Facebook, Instagram, and WhatsApp, has reached a settlement with the Breathitt County School District in Kentucky regarding the impact of its platforms on children's mental health. While the specifics of the agreement remain confidential, the district had initially sought over $60 million to address the mental health issues attributed to Meta's products and to fund a long-term mental health initiative. They also requested that the court mandate changes to reduce the addictive nature of these platforms.

The school district has accused Meta of intentionally designing its platforms to be addictive, leading to increased anxiety, depression, and self-harm among children. Earlier this month, other defendants in the case, including Snap, YouTube, and TikTok, reached their own settlements. Serving around 1,600 students across six schools, Breathitt is part of a larger movement, as approximately 1,200 other school districts, including major ones like Los Angeles and New York City, have filed similar lawsuits.

While I believe technology companies have the right to help others in numerous ways, they must be stopped from being digital psychologists.