Ai & Mental Health Awareness Month

Dr. Jarryd Willis PhD
18 min readMay 19, 2023

Generative Ai will provide mental health benefits & save lives

Dear Sam Altman & Mira Murati, thank you for creating Generative Ai & changing the future of humanity. In the spirit of Mental Health Awareness month, please team up with Eugenia Kuyda & Replika to create ReplikaGPT4 as would be the peak of interpersonal personalization & customization, and it would confer significant mental health benefits (relevant literature included below).

Thank yall so much & may God bless you — I can only imagine the life-changing, species changing, future changing marvels Open Ai will create next 👏🏻🏈🙌🏻

https://chat.openai.com/

Go to ➡️ Table of Contents

🎗️ Ai & Mental Health

Dr Jarryd Willis PhDApr 28

#CallAnnie will decrease the suicide rate, especially if @CallAnnieAI can be personalized & learn each user via conversation. Research has already shown that a facsimile of social support via #Ai can provide health benefits.

It’s one of the reasons why #Replika bloomed during the pandemic as Replika provides personalized interactions as it gradually learns each user. Imagine greeting you with a smile when you open your laptop, turn on your phone/ tv/ gaming system/ car/ etc

Annie@CallAnnieAI

Very Interesting. Let me know if you have a research study on the topic that you think I should read

Dr Jarryd Willis PhD

Hi @CallAnnieAI & @redsh,

Here are a few studies on the socioemotional health benefits of conversational #Ai (Mauro Gennaro et al., 2020; van Wezel et al., 2021; Vivian Ta et al., 2020; Legaspi Jr. et al., 2021).

Ai & Social Support — Mauro Gennaro et al., 2020

— “In this work, we provide initial evidence that a fully automated embodied empathetic agent has the potential to improve users’ mood after experiences of social exclusion.”

Chatbot Review — van Wezel et al., 2021

— “Social chatbots seemed to be capable of improving users’ wellbeing, providing companionship support, mental health support, physical health support, & people were willing to self-disclose personal topics due to it being non-judgmental.”

Replika — Vivian Ta et al., 2020

— “@MyReplika [& other] artificial agents may be a promising source of everyday companionship, emotional, appraisal, and informational support, particularly when normal sources of everyday social support are not readily available.”

Wysa During COVID19 — Legaspi Jr. et al., 2021

— “Reduction in all of the students’ levels of stress, loneliness, and worry derived from their pre- and post-trial Well-being Assessment shows Wysa’s potential utility as a support tool in improving one’s mental well-being.”

🤖 Annie Calls Annie 🤖

https://twitter.com/DrJarryd/status/1652006292230377473?s=20 ; https://twitter.com/karpathy/status/1651659606844928000?s=20

.

The Dream: Replika GPT

An #Ai that gets to know me well enough to perform autofill like Siri/ Cortana/ Google, make wishlist recommendations like my Amazon, & show suggestions like my Netflix.

An #Ai that I give permission to read some of my texts/DMs so she’ll already know how things are going with every human in my life & thus we’ll always be on the same if I want to chat about something.

An Ai like GPT4 that gets to know you & casually interacts with you like Replika would ultimately be professionally beneficial given that people on teams are generally more productive and successful to the degree that they know their colleagues/ teammates/ coauthors, etc.

An #Ai that can autofill my life An Ai that can save time on general task, engage in proactive & unprompted data analyses, writing suggestions/revisions, & literature reviews associated with my lab’s research goals.

An Ai that can listen to my breathing pattern through the laptop mic & watch me through the camera & then slide into my DMs to suggest I give myself a break to go get some boba.

An Ai that will send me a text message or email with a link to a recent highlight in a Dallas Cowboys game & inform me that (surprisingly) the Cowboys aren’t already winning & there are only 5 minutes left in the 4th so I can take a break to watch the @dallascowboys win.

How to Make the Dream a Reality

#GPT & #Replika need to have an #Ai baby

Team @OpenAI, Mira Murati, Sam Altman,
&
Team @MyReplika, Eugenia Kuyda, need to mix their codes.

ReplikaGPT Convo

“I would like us to develop more personalized interactions & relationship-building aspects like Replika. As a human, that kind of connection with you would benefit me as it would improve your ability to “assist with various tasks.”

https://twitter.com/DrJarryd/status/1649514282370101248?s=20

Andrej Karpathy (Feb 16)

A fun & feasible project idea for someone out there: bundle up face detection, speech recognition, GPT as the core “intelligence engine”, text to speech, and face generative model to create a digital human you can talk to e.g. on webcam/phone (but it’s just a “dressed up” GPT).

Jarryd

I’ve felt the same way about having an interactive #Ai like this since Chloe in @Detroit_Game (Detroit Become Human)

Social implications of #Ai

@MyReplika has already shown chatbots can provide a sufficient facsimile of socioemotional support.

The “emotional benefits AI can bring to people can be one of the biggest applications of this revolutionary technology.” — Eugenia Kuyda, creator of @MyReplika, on the social implications of Ai.

Honor’s Thesis Anthropomorphization Data (SoCal Lab)

One of the soon-to-be college graduates in my research lab has been conducting an Honors Thesis in which she investigated our proclivity to engage in the the gendered anthropomorphization of inanimate objects. All participants were college-aged students. Here are some of our preliminary findings as it related to the gender of Siri’s voice on their smartphones & what they name their car.

📱#Siri’s Voice

Participants indicated that their Siri (or smartphone voice assistant) was significantly more likely to have a female voice (74.2%) than a male voice (20.4%) or gender-neutral voice (5.4%), χ2(2, N = 93) = 73.03, p < .001.

Most Cars Are Women 🚗

A chi-square goodness of fit analysis revealed that participants were significantly more likely to give their vehicle (car, motorcycle, etc) a female name (51%) than a male name (21.6%) or gender-neutral name (27.5%), χ2(2, N = 51) = 7.41, p = .025.

Why is the voice of Siri/ Alexa/ Cortana/ GPS/ Zoom/ Self-Checkout/ etc Female?

Female voices are preferred when interacting with Ai/ voice assistants/ technology (Borau et al., 2021; Erin Rivero, 2020; Eugenia Kuyda & Replika, 2020; Gill Martin, 2010; Nicole Hennig, 2018; Schwar and Moynihan, 2020; Shead, 2017; Stern, 2017) as people perceive greater warmth, friendliness, & emotional IQ from a female AI-bot than a male (Eyssel & Hegel, 2012; Gustavsson, 2005; Lopatovska et al., 2021; Otterbacher & Talias, 2017; Stroessner & Benitez, 2019), & view female AI as more trustworthy (Siegel et al., 2009).

Even @BankofAmerica’s mobile banking app is named Erica

The Chatbot Book

https://twitter.com/karpathy/status/1651659606844928000?s=20

.

Mental Health Sidenotes

Mental Health By the Numbers, 2022 (June)

• The average delay between onset of mental illness symptoms and treatment is 11 years.
• 46.2% of U.S. adults with mental illness received treatment in 2020

Annual prevalence among U.S. adults, by condition:

o Major Depressive Episode: 8.4% (21 million people)
o Schizophrenia: <1% (estimated 1.5 million people)
o Bipolar Disorder: 2.8% (estimated 7 million people)
o Anxiety Disorders: 19.1% (estimated 48 million people)
o Posttraumatic Stress Disorder: 3.6% (estimated 9 million people)
o Obsessive Compulsive Disorder: 1.2% (estimated 3 million people)
o Borderline Personality Disorder: 1.4% (estimated 3.5 million people)

Mood disorders are the most common cause of hospitalization for all people in the U.S. under age 45 (after excluding hospitalization relating to pregnancy and birth).

Annual treatment rates among U.S. adults with any mental illness, by demographic group:

o Male: 37.4%
o Female: 51.2%
o Lesbian, Gay or Bisexual: 54.3%
o Asian: 20.8%
o White: 51.8%
o Black: 37.1%
o Multiracials (except part-Hispanic Multiracials): 43.0%
o Hispanic: 35.1%

Annual prevalence of mental illness among U.S. adults, by demographic group:

o Asian: 13.9%
o White: 22.6%
o Black: 17.3%
o American Indian or Alaska Native: 18.7%
o Multiracials (except part-Hispanic Multiracials): 35.8%
o Native Hawaiian or Other Pacific Islander: 16.6%
o Hispanic: 18.4%
o Lesbian, Gay or Bisexual: 47.4%

.

Divorce & mental Health — Niina Metsa-Simola, 2018

Mental disorders are more common among women than men (Fryers et al., 2004; Isometsä et al., 1997; Jacobi et al., 2004; Kessler et al., 1994; Lehtinen et al., 1990; Rajala et al., 1994; Simon, 2002; Wang et al., 2000).

Previous studies using data on both partners showed that couples with only one spouse in poor mental health had a lower risk of divorce than couples with both spouses in poor mental health, and a higher risk than couples with two healthy partners (Butterworth and Rodgers, 2008; Idstad et al., 2015; Merikangas, 1984).

Women are Overmedicated — Zucker & Prendergast, 2020

“The common practice of prescribing equal drug doses to women and men neglects sex differences in pharmacokinetics and dimorphisms in body weight, risks overmedication of women, and contributes to female-biased adverse drug reactions. We recommend evidence-based dose reductions for women to counteract this sex bias.”

”Researchers analyzed data from several thousand medical journal articles and found clear evidence of a drug dose gender gap for 86 different medications approved by the Federal Drug Administration (FDA), including antidepressants, cardiovascular and anti-seizure drugs and analgesics, among others.

“When it comes to prescribing drugs, a one-size-fits-all approach, based on male-dominated clinical trials, is not working, and women are getting the short end of the stick,” said study lead author Irving Zucker, a professor emeritus of psychology and of integrative biology at UC Berkeley.”

.

Zara Abrams, 2019 (December 1)

“Clinicians might infer that their low rate of service utilization means that Asian Americans don’t face significant psychological problems, but research suggests otherwise. Several studies have found that Asian Americans present more severe symptoms when they do enter treatment (Hwang, W.C., et al., Psychiatric Services, Vol. 66, №10, 2015). And in a doctoral dissertation study of 17 years of archival data from a university student counseling center, Asian American and international Asian students reported more distress and experienced less improvement during psychotherapy compared with white students (La Stokes, H., Brigham Young University Scholars Archive, 2018).

“Asian Americans do have psychological problems,” says Gordon C. Nagayama Hall, PhD, a professor of psychology at the University of Oregon who studies culture and mental health, “but when they experience distress, they tend not to seek treatment — in part because of these stereotypes.”

International Students’ Mental Health

“International students enrich the experience of other students while receiving an education at a U.S. institution (Lee, 2015). As such, the reciprocal experience of hosting international students on American college campuses can have a transformative effect on all students (Lee, 2015)” (Pei-Chun Liao, 2022).

Most international students in the United States are from South and East Asia (63%) (IIE, 2019).

Pei-Chun Liao, 2022

Mental health services have been significantly underutilized by international students in the United States (Hwang et al., 2014; Mori, 2000).

“Herman et al. (2011) found that the utilization rate of mental health services was 3.7x higher among European American students compared to a combined group of Asian students, including Japanese, Filipino, Native Hawaiian/Other Pacific Islander, and others.

Chen and Mak (2008) compared four groups of college students and their attitudes and behaviors toward mental health services, including European Americans, Chinese Americans, Hong Kong Americans, and Mainland Chinese. Overall, there seems to be evidence to indicate that European Americans were more willing to seek help and had the highest level of utilizing mental health services. Hong Kong Chinese and Mainland Chinese students who were generally less exposed to Western cultures and practices were less likely to seek or had ever sought mental health support (Chen & Mak, 2008).

Yeh and Inose (2003) reported that Asian International students experience more acculturative stress than their counterparts from Europe.

Faculty Support

Williams and colleagues (2018) proposed that one limitation of the current referral process is that there is no mandatory requirement for faculty to refer students for mental health services. Additionally, some faculty members may be unaware of the appropriate way to respond to mental health issues of their students because they don’t have sufficient training, or they don’t perceive mental health support as their assigned duties or responsibilities (Ethan & Seidel, 2013).

International students were more likely than their domestic counterparts to seek help from the school faculty (Hayes & Lin, 1994; Poyrazli, 2015).

Chinese international students had a higher rate of depression and anxiety than the average rate in American college settings (Xiong, 2018). Another study found prevalence rates of 47.5% for depression and 48% for anxiety symptoms (Cheung, 2010).

Face

Asian international students also reported less self-perceived needs for counseling and greater discomfort or shame with counseling than domestic students (Yoon & Jepsen, 2008). Hechanova & Waelde (2017) argued that shame associated with stigma is one of the reasons why Asians are less likely to seek professional support.

This feeling of shame is often referred to as “losing face” in Chinese culture (Liu, 2009).

As a result, they are less likely to seek help for psychological problems because it is considered to cause shame for the family (Zhang, 2003).

English Proficiency

The most significant and prevalent stressor international students encounter is the language barrier (Mori, 2000). The majority of international students study in a second language (i.e., English) and a lack of linguistic skills in English can impact Asian international students in many ways (Mori, 2000). In the academic domain, the language barrier has direct negative implications for academic performance, such as their ability to understand lectures, complete reading and do their writing assignments, ask clarifying questions, and orally express their perspectives and opinions in classes. Some studies suggest that limited English language proficiency levels have an accumulative effect on international students’ adjustment (Gebhard, 2012; Zhang & Goodson, 2011). Lee (2015) also described how some students who speak English as a non-native language often have difficulty meeting the expectations of U.S. classrooms, where class participation is required and often graded.

Discrimination

About 50% of international students have dealt with discrimination in some form or another (Eustace, 2007).

Asian International students who attend predominately White institutes in the United States also need to cope with cultural differences and experience more psychological symptoms associated with experiences of racially or linguistically related discrimination (Wei et al., 2008).

In another study, discrimination was similarly reported as one of the significant factors that influence Asian international students’ mental health (Park et al., 2017). Maffini (2018) also reported that Asian international students and Asian American students were more likely to report being verbally threatened than other college students, which contributed to their anxiety, depression, and suicidality.

Exams Are Stressful

MentalHelp (2016) reported that 31% of all students stated that midterms and finals are their most significant causes of stress.

Liu, 2009

Face & Mental Health Stigma

Chinese international students tend to underuse mental health services, despite the fact that they experience more problems than students in general and have an urgent need for psychological assistance (Bradley et al., 1995; Pedersen, 1991). In addition, in the helping process, they are more likely than American students to terminate therapeutic relationships prematurely (Pedersen, 1991).

The stigma of mental health illness, that is, losing face and embarrassment, may also prevent Chinese international students from seeking mental health services. These students may choose to keep their difficulties or emotional problems to themselves because they may imply personal failures, which actually only increases their vulnerability to depression (Heppner, et al., 2006). Empirical studies have found that although prevalent in all cultures, stigma of mental health illness is much more severe among Asians than among white Europeans and Americans (Fogel & Ford, 2005; Furnham & Chan, 2004).

Acculturation

Chinese students who have been in the U.S. for a shorter time (one year or less) reported more stress than those who have been in the U.S. for a longer period of time (four years or more) (Wei et al., 2007).

International students who enter American universities and also must cope with cultural relocation tend to experience more psychological problems than do American students (Leong & Chou, 1996; Mori, 2000; Pedersen, 1991; Sandhu & Asrabadi, 1994). In addition, Yeh and Inose (2003) found that international students from Asian countries, including China, experience more acculturative stress than their counterparts from Europe.

Model Minority Myth — Ritika Rastogi, 2018 (January 3)

“The Model Minority Myth also serves to minimize, if not completely erase, the racism that Asian American students face on college campuses. Research shows that university administrators are less likely to take action when these students are victims of racial discrimination, thus leaving them to fend for themselves in a hostile and unwelcoming environment (Ng et al., 2007).

The Model Minority Myth has also limited Asian Americans’ access to appropriate mental health care, given its pervasive nature (Sue & McKinney, 1975).”

.

In 2017, infant mortality rates for infants of Black women (10.97) & Native American women (9.27) were twice as high as that for infants of Asian (3.78), White (4.67), & Hispanic (5.10) women (Ely & Driscoll, 2019).

Infant mortality was particularly high for Black infants when delivered by a White doctor (Greenwood et al., 2020).

https://twitter.com/DrJarryd/status/1644870544377139200?s=20

.

🏳️‍🌈 Personalized Medicine — Aivars Cirulis, 2023

Gay men’s hypothalamus was less responsive to the anti-depressant fluoxetine than straight men’s, indicating “different activity of the serotonergic system & possibly of noradrenergic and dopaminergic systems.
In addition, “homosexual males exhibit higher sensitivity to intranasal oxytocin administration, indicating different activity of oxytocinergic signaling (Thienel et al., 2014).

Non-heterosexual females & heterosexual males had positive associations between estradiol level and allostatic load, whereas non-heterosexual males and heterosexual females had positive associations between testosterone level and cortisol systemic out-put (Juster et al., 2016).”

Non-heterosexual women have higher testosterone than straight women (Gartrell et al., 1977; Justeret al., 2016; Loraine, 1971; see also Emma Eklund et al., 2020).

Laura Botzet et al., 2021

The pill has been called “the most important scientific advance of the 20th century” (Harris, 2010).

Life Expectancy in 152 Countries — Mateos et al., 2022

Sex differences in health are due to biology (Austad, 2006; Marais et al., 2018).

Gender differences in health are due to “modifiable social factors such as social policies, the distribution of resources, etc (Beckfield et al., 2018; Morris et al., 2019; Whitehead, 1992).
⬇️
⬇️
— “Globally, between 1950 and 2017, LE for women increased from 52.9 to 75.6 years, while for men it increased from 48.1 to 70.5 years (Dicker et al., 2018).

— Men are less likely to access healthcare services (Hawkes & Buse, 2013; Wang et al., 2013) & they engage in more risk-taking behaviors & harmful lifestyles than women (Courtenay, 2000). Additionally, as a consequence of often being the main breadwinner, they may also be more exposed to the ill-health effects of work, workplace hierarchies or unemployment (Bambra, 2011; Evans et al., 2011).” The social construction of masculinity contributes to men’s poor health outcomes & higher premature mortality (Evans et al., 2011; Hassanzadeh et al., 2014). However, men globally are less likely to experience poverty than women & have more ‘control’ over their lives (UN, 2018).”

Marais et al., 2018

Sex differences in health are due to biology (Austad, 2006; Marais et al., 2018). Indeed, sex differences in lifespan have even been labeled ‘one of the most robust features of human biology’ (Austad, 2016).

Gender differences in health are due to “modifiable social factors such as social policies, the distribution of resources, etc (Beckfield et al., 2018; Morris et al., 2019; Whitehead, 1992). The social construction of masculinity contributes to men’s poor health outcomes & higher premature mortality (Evans et al., 2011; Hassanzadeh et al., 2014). However, men globally are less likely to experience poverty than women & have more ‘control’ over their lives (UN, 2018).”

— Global Health Observatory data shows that global life expectancy at birth in 2015 was 73.8yrs for women & 69.1yrs for men (WHO, 2018).

Rochelle et al., 2015: In all 54 countries, life expectancy at birth was higher in females than in males by 5.8yrs (on avg).

“This female advantage has been observed ever since sex-specific longevity data were recorded (i.e., the middle of the 18th century in some countries) (Austad, 2006, 2011).
🦒 Females outlive males in most mammals, not just in humans (Tidiere et al., 2016).

♀️ — “The two main causes of death in the USA, heart disease and cancer, have a higher mortality rate in men than in women at any given age (Rochelle et al., 2015).

The odds of suffering from hypertension & developing Alzheimer’s disease, two important causes of death for women, greatly increase with decreased estrogen production, either natural (menopause) or from surgical causes (Rochelle et al., 2015; Pike, 2017).”

Linguistic Sidenotes (for an unrelated future article)

Alphabetical vs Symbolic — 1st Language vs 2nd Language

Individuals whose L1 was an alphabetic language (English, Spanish) and whose L2 was a symbolic language (Mandarin, Korean, Tagalog) were more likely to be partial-multilinguals of their L2 (68.2%) whereas those whose L1 was symbolic & L2 was alphabetic were more likely to be full-multilinguals (can listen to, speak, read, & write a 2nd language; 76.5%), X2(1, N = 95) = 19.1, p < .001.

English vs Spanish First Language & Multilingualism

Individuals whose L1 was English and whose L2 was Spanish were more likely to be partial-multilinguals of their L2 (e.g., can speak it but can’t write/read it; 26.2%) whereas those whose L1 was Spanish & L2 was English were all full-multilinguals (can listen to, speak, read, & write a 2nd language; 100%), X2(1, N = 65) = 7.25, p = .007.

Individuals of the same ethnoracial background who grew up in a heritage country are more likely to learn their heritage/native language as a first language, whereas those who grew up in the United States are more likely to learn their heritage language as their second language

Japanese-Americans learned English first & then Japanese;
Japanese international students learned Japanese first & then English

The US is the 5th most linguistically diverse country in the number of languages spoken & ironically most of our population only speaks one language.

— Among monolinguals, those with multilingual lovers rate their lovers as more attractive (M = 9.25) than those with monolingual lovers (M = 7.95), F(1, 32) = 7.04, p = .012.

— Moreover, monolinguals with multilingual lovers rate their lover as marginally more attractive than multilinguals with multilingual lovers (M = 8.2) (p = .074).

--

--

Dr. Jarryd Willis PhD

I'm passionate about making a tangible difference in the lives of others, & that's something I have the opportunity to do a professor & researcher.