OpenAi's new ChatGPT

Wits_End said:
If I've understood correctly how GPs are trained in decision-making (at least in the UK) these days, it does sound rather as if they use a flowchart-type principle:

Patient presents with symptom X:
>
Check temperature. Is it normal? Y/N
>
If Yes, do A. If No, do B
>
and so on. So is there any reason why AI couldn't do the same?
Click to expand...
Yup. AI will be massively better at it because it'll be as if an entire team of all the specialties had ample time to consider every possibility.

It will even do the empathy part a lot better, because the source for its knowledge base will be the patients themselves.

It will probably be the single most revolutionary technology in the history of medicine. One of the biggest issues with healthcare everywhere is shortages. There is a shortage of everyone.

I saw something yesterday about a clinic in a rural city in Canada who couldn't even find a GP to staff a clinic despite a $200K cash bonus. There aren't enough doctors and they don't have enough time to do half the work they're supposed to do.

Expert time is the most valuable resource here. And the system is built on shortage of all of it, the system "works" by neglecting over half of what needs to be done. Time for research, for learning, for listening, for deciding.

There will always be a need for human physicians, AIs won't always work everywhere all the time. There will be a need for offline expertise. But what's coming will be better, faster and cheaper in ways that only information technology can bring: not a little bit, not some 10-15% or so, but 10-100x.
 
one part of the NHS where I can see it being introduced is the 111 service. Certainly the front end, sort of triage, bit is very much a 'follow the script' exercise which could easily be done by AI and then if necessary put through to a clinician (which is what already happens now).
 
Has anyone asked it about the name of our disease? If it had to rename ME, what new name would it suggest?
 
mango said:
Has anyone asked it about the name of our disease? If it had to rename ME, what new name would it suggest?
Click to expand...
  1. Chronic Exhaustion Syndrome (CES)
  2. Chronic Fatigue Impairment Disorder (CFID)
  3. Post-Exertional Malaise Syndrome (PEMS)
  4. Immune Deficiency Fatigue Syndrome (IDFS)
  5. Neurogenic Exhaustion Syndrome (NES)
  6. Systemic Energy Conservation Disorder (SECD)
  7. Neuro-Immune Exhaustion Syndrome (NIES)
  8. Endurance Limitation Syndrome (ELS)
  9. Neurological Debilitation Syndrome (NDS)
  10. Chronic Energy Deficiency Syndrome (CEDS)
  1. Systemic Energy Imbalance Disorder (SEID)
  2. Chronic Energy Deficiency Disorder (CEDD)
  3. Systemic Activity Limitation Syndrome (SALS)
  4. Progressive Exhaustion Disorder (PED)
  5. Chronic Exercise Intolerance Syndrome (CEIS)
  6. Systemic Limitation Disorder (SLD)
  7. Pervasive Debilitation Syndrome (PDS)
  8. Chronic Fatigue Inhibitor Disorder (CFID)
  9. Generalized Fatigue Condition (GFC)
  10. Chronic Energy Management Disorder (CEMD)
I did try to get it not to use 'fatigue' or 'syndrome' but it did anyway.

eta:
these were better
  1. Chronic Neuroimmune Disease
  2. Post-Exertional Neurological Disorder
  3. Systemic Exertion Intolerance Disease
  4. Neuroendocrine Immune Disease
  5. Severe Chronic Immune Dysfunction
  6. Malaise-Neuroinflammatory Disease
  7. Immune-Mediated Neurological Disease
  8. Chronic Multisystemic Inflammatory Disease
  9. Myalgic Neuroimmunodeficiency Disorder
  10. Chronic Post-Infectious Neuroinflammatory Disorder.
 
Last edited:
Sly Saint said:
  1. Chronic Exhaustion Syndrome (CES)
  2. Chronic Fatigue Impairment Disorder (CFID)
  3. Post-Exertional Malaise Syndrome (PEMS)
  4. Immune Deficiency Fatigue Syndrome (IDFS)
  5. Neurogenic Exhaustion Syndrome (NES)
  6. Systemic Energy Conservation Disorder (SECD)
  7. Neuro-Immune Exhaustion Syndrome (NIES)
  8. Endurance Limitation Syndrome (ELS)
  9. Neurological Debilitation Syndrome (NDS)
  10. Chronic Energy Deficiency Syndrome (CEDS)
  1. Systemic Energy Imbalance Disorder (SEID)
  2. Chronic Energy Deficiency Disorder (CEDD)
  3. Systemic Activity Limitation Syndrome (SALS)
  4. Progressive Exhaustion Disorder (PED)
  5. Chronic Exercise Intolerance Syndrome (CEIS)
  6. Systemic Limitation Disorder (SLD)
  7. Pervasive Debilitation Syndrome (PDS)
  8. Chronic Fatigue Inhibitor Disorder (CFID)
  9. Generalized Fatigue Condition (GFC)
  10. Chronic Energy Management Disorder (CEMD)
I did try to get it not to use 'fatigue' or 'syndrome' but it did anyway.

eta:
these were better
  1. Chronic Neuroimmune Disease
  2. Post-Exertional Neurological Disorder
  3. Systemic Exertion Intolerance Disease
  4. Neuroendocrine Immune Disease
  5. Severe Chronic Immune Dysfunction
  6. Malaise-Neuroinflammatory Disease
  7. Immune-Mediated Neurological Disease
  8. Chronic Multisystemic Inflammatory Disease
  9. Myalgic Neuroimmunodeficiency Disorder
  10. Chronic Post-Infectious Neuroinflammatory Disorder.
Click to expand...
Very interesting :) Some of them are actually quite good, I think? Thank you so much, I really appreciate it :thumbsup:
 
I wish I could say that it's shocking that AIs are massively overperforming human doctors at empathy, but I pretty much expected it to be like this. It's basically like a mirror of each other. This is the difference that learning straight from those with lived experience does, removing the massive bias and hubris that take hold instead.

But seriously, wow. The better quality is expected, but for robot intelligence to be more human than actual humans is just as shocking as it's expected:

Fuzv0dkaUAE_4Ta


 
mango said:
AIs are data/compute systems. AIs are not capable of empathy.
Click to expand...

I think that is the point of such comments. Even forty plus years ago, patients reported automated computerised psychiatric interviews responding with repetition and to key word triggers as being experienced as more understanding and empathetic than actual professionals. So it is hardly surprising that more recent AI systems are perceived by the patient as more ‘empathetic’.
 
Peter Trewhitt said:
I think that is the point of such comments. Even forty plus years ago, patients reported automated computerised psychiatric interviews responding with repetition and to key word triggers as being experienced as more understanding and empathetic than actual professionals. So it is hardly surprising that more recent AI systems are perceived by the patient as more ‘empathetic’.
Click to expand...
I agree, Peter.

And I believe that it's important that we choose our words carefully when discussing these things. Let's not even for a second pretend that the AI output is somehow comparable to genuine human empathy. They are two completely different things. It's not empathy, so let's not call it empathy.

I can't see how we can have the meaningful conversations that are needed, as long as people (media, social media, researchers, everyone everywhere...) keep wrongly attributing human traits like emotions or intentions to AI. It's just misleading and takes focus away from the real stuff. It really irks me.

I agree that the study's findings are interesting and worth discussing, but I feel that the whole conversation would be meaningless if we don't first make sure to call things by their real names. It's not empathy, so please let's not call it empathy.

(I'm a big fan of works of fiction such as the movies Zoe and Her, as well as Martha Wells' books about Murderbot. But again, that's a completely different thing.)
 
Last edited:
Stochastic parrots
At its core, the term “stochastic parrots” refers to large language models that are impressive in their ability to generate realistic-sounding language but ultimately do not truly understand the meaning of the language they are processing.
_________________

I heard an expert on AI describing it as a stochastic parrot, and therefore just as likely to produce misinformation as good information and unable to tell the difference.
 
Trish said:
Stochastic parrots
At its core, the term “stochastic parrots” refers to large language models that are impressive in their ability to generate realistic-sounding language but ultimately do not truly understand the meaning of the language they are processing.
_________________

I heard an expert on AI describing it as a stochastic parrot, and therefore just as likely to produce misinformation as good information and unable to tell the difference.
Click to expand...
How... human.
 
Trish said:
Stochastic parrots
At its core, the term “stochastic parrots” refers to large language models that are impressive in their ability to generate realistic-sounding language but ultimately do not truly understand the meaning of the language they are processing.
_________________

I heard an expert on AI describing it as a stochastic parrot, and therefore just as likely to produce misinformation as good information and unable to tell the difference.
Click to expand...
Good example. That term was used in this paper: https://dl.acm.org/doi/pdf/10.1145/3442188.3445922.

Here's a statement from those same authors that's much easier to read.


There's even a short TL;DR ("too long; didn't read") section at the top:
The harms from so-called AI are real and present and follow from the acts of people and corporations deploying automated systems. Regulatory efforts should focus on transparency, accountability and preventing exploitative labor practices.
Click to expand...

One thing that helps with transparency is making sure the language used does not lead folks to anthropomorphize these programs. On one level we may be aware that these things are not sentient. They don't have consciousness. But we may subconsciously assume a mind or intention due to the language that is used when they answer questions.

For example, some experts have recommended using third person language rather than I statements. Here's one example.

Screen Shot 2023-05-29 at 8.37.19 AM.png
Click to expand...

I thought this example was so interesting because the first answer seems just fine until you read the second answer.
 
A thought experiment by Emily Bender.

This is a post on Medium, and I wasn't sure the embedded version worked, so I've also included a "code" link.
Code: 
https://medium.com/@emilymenonbender/thought-experiment-in-the-national-library-of-thailand-f2bf761a8a83

 
My latest test of ChatGPT:

Write a short article for a wellness website about Chronic Fatigue Syndrome, including symptoms, diagnosis, causes and treatment.

Title: Understanding Chronic Fatigue Syndrome: Symptoms, Diagnosis, Causes, and Treatment

Introduction: Chronic Fatigue Syndrome (CFS), also known as myalgic encephalomyelitis (ME), is a complex and debilitating condition that affects millions of people worldwide. Characterized by persistent fatigue and a range of other symptoms, CFS can significantly impact an individual's daily life and overall well-being. In this article, we will explore the symptoms, diagnosis, causes, and treatment options associated with Chronic Fatigue Syndrome.

Symptoms: The primary symptom of CFS is excessive fatigue that is not relieved by rest or sleep. However, individuals with CFS experience a wide array of additional symptoms, which may include:

  1. Severe fatigue lasting for at least six months
  2. Unrefreshing sleep
  3. Cognitive difficulties, often referred to as "brain fog"
  4. Muscle and joint pain
  5. Sore throat
  6. Headaches
  7. Swollen lymph nodes
  8. Post-exertional malaise, where symptoms worsen after physical or mental exertion
  9. Unexplained muscle weakness
  10. Sensitivity to light and sound
Diagnosis: Diagnosing CFS can be challenging due to the absence of specific laboratory tests or biomarkers. Healthcare professionals typically rely on a thorough medical history assessment, physical examination, and the exclusion of other medical conditions with similar symptoms. The diagnostic process often involves ruling out conditions such as sleep disorders, autoimmune diseases, and mental health disorders.

Causes: The exact causes of CFS are still not fully understood. However, researchers believe that it may result from a combination of various factors, including:

  1. Viral infections: Some cases of CFS have been linked to viral infections like Epstein-Barr virus (EBV), human herpesvirus 6 (HHV-6), and others.
  2. Immune system dysfunction: Abnormalities in the immune system may contribute to the development of CFS.
  3. Hormonal imbalances: Disruptions in hormone levels, particularly cortisol and serotonin, have been observed in individuals with CFS.
  4. Genetic predisposition: Certain genetic factors might make individuals more susceptible to developing CFS.
  5. Environmental factors: Exposure to toxins and stressful life events may trigger or worsen symptoms in susceptible individuals.
Treatment: At present, there is no cure for Chronic Fatigue Syndrome. However, various approaches can help manage symptoms and improve quality of life:

  1. Lifestyle modifications: Prioritizing rest, pacing activities, and avoiding overexertion are vital in managing symptoms.
  2. Cognitive Behavioral Therapy (CBT): This therapeutic approach helps individuals develop coping strategies, address negative thought patterns, and manage stress.
  3. Graded Exercise Therapy (GET): Under the guidance of a healthcare professional, GET involves gradually increasing physical activity levels to build tolerance and reduce symptoms.
  4. Medications: Depending on the symptoms, healthcare providers may prescribe medications to alleviate pain, improve sleep, or manage other specific symptoms.
  5. Supportive therapies: Complementary and alternative therapies such as acupuncture, massage, and relaxation techniques may provide some relief for certain individuals.
Conclusion: Chronic Fatigue Syndrome is a complex condition that can significantly impact a person's daily life and well-being. While the exact causes remain unclear, individuals with CFS can work with healthcare professionals to manage their symptoms through a combination of lifestyle modifications, therapeutic interventions, and supportive therapies. Raising awareness about CFS and providing appropriate support can help individuals navigate this challenging condition and improve their overall quality of life.
 
Then I gave it much the same question for a different context:

Write a short article for a university medical department website about chronic fatigue syndrome. Include symptoms, diagnosis, causes and treatment.

Title: Chronic Fatigue Syndrome: An Overview of Symptoms, Diagnosis, Causes, and Treatment

Introduction: Chronic Fatigue Syndrome (CFS), also known as myalgic encephalomyelitis (ME), is a complex and debilitating condition that affects individuals, including many university students. Understanding its symptoms, diagnosis, causes, and treatment is crucial for medical professionals and students alike. This article provides an overview of CFS, shedding light on its various aspects.

Symptoms: The primary symptom of CFS is persistent fatigue that lasts for at least six months and is not alleviated by rest or sleep. However, CFS encompasses a wide range of additional symptoms, which may include:

  1. Impaired cognitive function: Individuals with CFS often experience difficulties with memory, concentration, and overall mental clarity, often referred to as "brain fog."
  2. Muscle and joint pain: Widespread pain and discomfort in the muscles and joints are common among CFS patients.
  3. Unrefreshing sleep: Despite adequate sleep, individuals with CFS wake up feeling unrefreshed, which further contributes to their overall fatigue.
  4. Post-exertional malaise (PEM): Physical or mental exertion can lead to a worsening of symptoms, causing prolonged fatigue and increased pain.
  5. Sore throat and tender lymph nodes: Many individuals with CFS experience a persistent sore throat and swollen or tender lymph nodes.
  6. Headaches: Frequent headaches, often resembling migraines, are commonly reported by CFS patients.
  7. Sensitivity to light and sound: Individuals with CFS may become hypersensitive to bright lights, loud noises, or certain smells.
Diagnosis: Diagnosing CFS can be challenging due to the absence of specific laboratory tests or biomarkers. Healthcare professionals must rely on a comprehensive approach, which typically involves:

  1. Medical history assessment: A thorough evaluation of the patient's medical history and symptoms is necessary to identify patterns consistent with CFS.
  2. Physical examination: Conducting a physical examination helps rule out other medical conditions that may cause similar symptoms.
  3. Exclusion of alternative diagnoses: To eliminate other possible causes, healthcare providers will consider a range of conditions, such as sleep disorders, autoimmune diseases, and mental health disorders.
  4. Symptom monitoring: Keeping track of symptoms over time helps identify patterns and evaluate their impact on daily life.
Causes: The precise causes of CFS are not yet fully understood. However, research suggests that CFS may result from a combination of factors, including:

  1. Viral infections: Some cases of CFS have been associated with viral infections like Epstein-Barr virus (EBV) and human herpesvirus 6 (HHV-6).
  2. Immunological dysfunction: Abnormalities in the immune system may play a role in the development of CFS.
  3. Neuroendocrine imbalances: Disruptions in hormone levels, particularly cortisol and serotonin, have been observed in individuals with CFS.
  4. Genetic predisposition: Certain genetic factors might make individuals more susceptible to developing CFS.
  5. Environmental triggers: Exposure to toxins and stressful life events may act as triggers or exacerbate symptoms in susceptible individuals.
Treatment: While there is currently no cure for CFS, various treatment approaches can help manage symptoms and improve quality of life. These may include:

  1. Lifestyle modifications: Emphasizing the importance of rest, pacing activities, and avoiding overexertion is crucial in managing CFS symptoms.
  2. Cognitive Behavioral Therapy (CBT): CBT can assist individuals in developing coping mechanisms, addressing negative thoughts, and managing stress associated with CFS.
  3. Graded Exercise Therapy (GET): Gradually increasing physical activity levels under professional guidance can help build tolerance and reduce symptoms.
  4. Medications: Depending on the symptoms and individual needs, healthcare providers may prescribe medications to alleviate pain, improve sleep, or manage specific symptoms.
  5. Supportive therapies: Complementary and alternative therapies, such as acupuncture, massage, and relaxation techniques, may provide some relief and support for certain individuals.
Conclusion: Chronic Fatigue Syndrome is a challenging condition that affects individuals, including university students. Recognizing the symptoms, diagnosing the condition accurately, understanding potential causes, and implementing appropriate treatment strategies are crucial for medical professionals and students alike. By raising awareness and providing effective support, the medical community can contribute to improving the management and overall well-being of individuals living with CFS.
 
You must log in or register to reply here.
Back
Top Bottom