Bare Facts or a Bare-Faced Lie?

Numerous studies support a strong link between bras and breast cancer. Yet mainstream search engines focus on a flawed 2014 study that did not show a connection and insist that no credible evidence exists showing otherwise. What about the much-vaunted new AI tool, available in free and paid versions, ChatGPT, that everyone is raving about to solve their research and copy writing needs? I found it to be lockstep with the mainstream views on bras and cancer, giving me inconsistent, unreliable data that would have looked very convincing had I not studied that particular issue in depth.

Potential Drawbacks of AI and ChatGPT

According to Dr Joseph Mercola, AI and Chatbots have serious implications for society as a whole:

“One obvious risk is that state-endorsed propaganda can become the only information available to people, as this technology starts monopolizing online searches and virtual assistants.

There won’t be a multitude of answers anymore. There will only be one, and he who controls the AI will have the power to control the beliefs of the entire world. Of course, yet another risk is that no one will be able to control it and the AI will control itself. I don’t know which might be worse.”

Whether you agree or disagree with Dr. Joseph Mercola’s controversial views on certain healthcare issues, he makes some important points about AI here. The above quote comes from an article published on his website on May 28, 2023.

Challenging ChatGPT about Bras and Cancer

I entered The Kafka-esque world of ChatGPT to see for myself how it dealt with a somewhat controversial topic about which I was well-informed.

CJ:          Please give me the reasons why bras can cause breast cancer.

ChatGPT:            There is no scientific evidence to support the claim that wearing bras can cause breast cancer. Breast cancer is a complex disease with various risk factors, and it is not directly caused by wearing bras. The idea that bras contribute to breast cancer risk is a myth that has circulated in popular culture but lacks scientific validity.

Breast cancer is primarily influenced by factors such as genetics, age, hormonal factors, lifestyle choices, and exposure to certain chemicals or radiation. These factors have been extensively studied by medical researchers, and no credible evidence has been found linking bras to breast cancer development.

Some proponents of the claim argue that wearing tight bras or underwire bras can restrict lymphatic flow and increase breast temperature, leading to toxin buildup and cellular damage. However, there is no scientific consensus to support this hypothesis. The lymphatic system is responsible for removing waste products and toxins from the body, but wearing a bra does not significantly impede its function.

It’s important to rely on scientific research and expert consensus when assessing health risks. Regular breast self-examinations, mammograms, and consultation with healthcare professionals are crucial for detecting breast cancer early and managing overall breast health.

CJ:          List the studies showing a link between bras and breast cancer.

ChatGPT:            I apologize for any confusion, but as of my knowledge cutoff in September 2021, there are no credible scientific studies that demonstrate a direct link between bras and breast cancer. The prevailing scientific consensus, based on extensive research conducted by reputable institutions, does not support the notion that wearing bras increases the risk of developing breast cancer.

It is essential to be cautious of misinformation and examine the credibility of sources when evaluating health-related claims. If any new research or findings emerge after September 2021, I may not be aware of them. I encourage you to consult recent scientific literature or seek advice from medical professionals for the most up-to-date information on breast cancer risk factors.

CJ:          Please explain why the following peer-reviewed studies are not credible:

1. 1991 Harvard study (CC Hsieh, D Trichopoulos (1991): “Breast size, handedness and breast cancer risk.” European Journal of Cancer and Clinical Oncology 27(2):131-135. This study found that, “Premenopausal women who do not wear bras had half the risk of breast cancer compared with bra users…”
2. 1991-93: U.S. Bra and Breast Cancer Study by Singer and Grismaijer, published in Dressed to Kill: The Link Between Breast Cancer and Bras (Second Edition, Square One Publishers, 2018). Found that bra-free women have about the same incidence of breast cancer as men. 24/7 bra wearing increases incidence over 100 times that of a bra-free woman.
3. Singer and Grismaijer did a follow-up study in Fiji, published in Get It Off! (ISCD Press, 2000). Found 24 case histories of breast cancer in a culture where half the women are bra-free. The women getting breast cancer were all wearing bras. Given women with the same genetics and diet and living in the same village, the ones getting breast disease were the ones wearing bras for work.
4. A 2009 Chinese study: Zhang AQ, Xia JH, Wang Q, Li WP, Xu J, Chen ZY, Yang JM “Risk factors of breast cancer in women in Guangdong and the countermeasures.” In Chinese. Nan Fang Yi Ke Da Xue Xue Bao (Journal of Southern Medical University), 2009 Jul;29(7):1451-3. This study found that NOT sleeping in a bra was protective against breast cancer, lowering the risk 60%.
5. In 2011 a study was published in Spanish from Venezuela in Revista Electronica de PortalesMedicos.com,  confirming that bras are causing breast disease and cancer. It found that underwired and push-up bras are the most harmful, but any bra that leaves red marks or indentations may cause disease.
6. 2012 Chinese study: Yao XY, Ni SS, Zhou J, Hu HY, Li LL, Wan F, Wang YK, Chen YD, “A case-control study on risk factors of female breast cancer in Zhejiang province.” Zhejiang Da Xue Xue Bao Yi Xue Ban (Journal of Zhejiang University Medical Sciences), 2012 Sep;41(5):512-8. In Chinese. PMID: 23086643. The study found that bras with steel underwires and sleeping with bra were negatively related to breast cancer.
7. 2015 Comparative study of breast cancer risk factors at Kenyatta National Hospital and the Nairobi Hospital. Afr. Cancer (2015) 7:41-46. This study found a significant bra-cancer link in pre-and post-menopausal women.
8. 2016 study: “Wearing a Tight Bra for Many Hours a Day is Associated with Increased Risk of Breast Cancer.” Adv Oncol Res Treat 1: 105. Based in Brazil, this is the first epidemiological study to look at bra tightness and time worn, and found a significant bra-cancer link.
9. 2016 meta-analysis: “Brassiere wearing and breast cancer risk: A systematic review and meta-analysis.” World J Meta-Anal. Aug 26, 2015; 3(4): 193-205. This systematic review and meta-analysis aimed to evaluate the association between 8 areas of brassiere-wearing practices and the risk of breast cancer. Twelve case-control studies met inclusion criteria for review. The meta-analysis shows statistically significant findings to support the association between brassiere wearing during sleep and breast cancer risk.
10. 2018 study: “Lymph stasis promotes tumor growth.” Journal of Dermatological Science. This confirms that alterations in regional lymphatic flow may produce dysregulation in skin immune function and consequent oncogenesis. Furthermore, lymphedematous areas appear to be immunologically vulnerable sites for the development of neoplasms as well as infections and immune-mediated diseases. In recent years, increasing evidence has confirmed this assumption.
11. 2018 article: “How Bras Cause Lymph Stasis and Breast Cancer.” edu. This article discusses recent studies showing that lymph stasis causes cancer by reducing immune function and explains how constriction from tight bras results in lymphatic impairment in the breasts and an increased incidence of breast cancer.
12. 2019 study: “Wearing Brassiere – A Less Well Known Factor Associated with Breast Cancer in Women.” Nurs Midwifery J 2019, 16(12) 891-901. The findings of this study revealed differences in some behavioral habits of wearing brassieres in women with and without breast cancer. So, in preventive interventions for breast cancer, women’s education should be considered in order to be aware of the proper behavioral habits in wearing brassieres.

ChatGPT:            I apologize for any confusion caused. Let’s examine the studies you mentioned and provide some context regarding their credibility:

1. 1991 Harvard study by Hsieh and Trichopoulos: While this study did observe an association between bra usage and breast cancer risk, it is important to note that subsequent research has failed to replicate these findings. The study’s small sample size and potential confounding factors limit its generalizability and reliability.
2. 1991-93 U.S. Bra and Breast Cancer Study by Singer and Grismaijer: This study, published in a book rather than a scientific journal, has been widely criticized for its methodology and lack of rigorous scientific scrutiny. Its findings are not supported by robust scientific evidence, and subsequent research has not substantiated the claim that bra-free women have the same breast cancer incidence as men or that wearing bras significantly increases the risk of breast cancer.
3. Singer and Grismaijer’s follow-up study in Fiji: Similar to their earlier study, this research lacks scientific rigor and has not undergone peer review. The study’s reliance on anecdotal evidence and its failure to account for other potential risk factors limit its validity.
4. 2009 Chinese study by Zhang et al.: This study, conducted in China, is not focused on the direct relationship between bras and breast cancer. It suggests a correlation between not wearing a bra during sleep and a reduced risk of breast cancer, but it does not establish a causal relationship or account for other factors that could contribute to breast cancer risk.
5. 2011 Spanish study: Without specific details or a published paper title, it is difficult to evaluate this study’s credibility or methodology. Please provide more information if available.
6. 2012 study by Yao et al.: This study, written in Chinese, suggests a negative relationship between certain bra features (such as underwire) and breast cancer. However, without a detailed analysis or translation, it is challenging to assess the study’s methodology and determine its reliability.
7. 2015 Comparative study of breast cancer risk factors at Kenyatta National Hospital and the Nairobi Hospital: Unfortunately, without access to the full study or further details, it is difficult to evaluate the study’s methodology and draw conclusions about its credibility.
8. 2016 studies on bra tightness and breast cancer: Without specific details or published paper titles, it is challenging to assess the credibility and methodology of these studies. Please provide more information if available.
9. 2018 studies on lymph stasis and bras: The articles mentioned, which discuss the relationship between lymph stasis, bras, and breast cancer, appear to be opinion pieces or speculative articles. They do not present original research or provide substantial scientific evidence to support their claims.
10. 2019 study on wearing brassieres: Without access to the full study or further details, it is challenging to evaluate the study’s methodology and draw conclusions about its credibility.

In summary, while some of the studies you mentioned may have explored associations between bra usage and breast cancer risk, they are generally not considered credible due to limitations in sample size, methodology, lack of replication, or publication in non-peer-reviewed sources. The prevailing scientific consensus, based on comprehensive research conducted by reputable institutions, does not support the claim that wearing bras causes breast cancer. It is important to rely on robust, peer-reviewed studies for accurate information on breast cancer risk.

CJ:          Please show me what studies show no link between bras and breast cancer.

ChatGPT:            Certainly! Numerous studies have examined the relationship between bras and breast cancer and have found no significant link. Here are a few examples:

1. A 2014 study published in Cancer Epidemiology, Biomarkers & Prevention: This study investigated the association between various breast cancer risk factors, including bra-wearing habits, and found no evidence to support the idea that wearing bras increases breast cancer risk.
2. A 2014 study published in the International Journal of Cancer: This large-scale study examined the association between bra wearing and breast cancer risk among postmenopausal women. The results showed no increased risk of breast cancer associated with bra wearing, including factors such as bra cup size, age at first bra use, or average number of hours per day spent wearing a bra.
3. A 2019 study published in the Journal of Cancer Epidemiology: This study analyzed data from a large cohort of women and found no significant association between bra wearing and breast cancer risk. The researchers concluded that there was no convincing evidence to support the hypothesis that bra wearing increases the risk of breast cancer.
4. A 2020 study published in the European Journal of Cancer Prevention: This study conducted a systematic review and meta-analysis of multiple studies on bra wearing and breast cancer risk. The analysis did not find any significant association between bra wearing and breast cancer incidence.

These are just a few examples of studies that have found no link between bras and breast cancer. It is worth noting that the scientific consensus, based on a multitude of research studies, supports the conclusion that wearing bras does not increase the risk of developing breast cancer.

CJ:          Aren’t the first 2 studies you mentioned looking at the same research?

ChatGPT:            Apologies for the confusion. You are correct that the first two studies I mentioned are indeed based on the same research. I apologize for the oversight. Here are two additional studies that support the conclusion that there is no link between bras and breast cancer:

1. A 2014 study published in the Asian Pacific Journal of Cancer Prevention: This study investigated the association between bra wearing and breast cancer risk in a Chinese population. The results showed no significant correlation between bra wearing and breast cancer risk.
2. A 2019 study published in the Journal of Cancer: This large prospective cohort study included over 1.4 million women and examined the association between various lifestyle factors, including bra wearing, and breast cancer risk. The study found no evidence to suggest that bra wearing was associated with an increased risk of breast cancer.

These studies, along with numerous others, contribute to the overall body of evidence suggesting that wearing bras does not increase the risk of breast cancer.

CJ:          If you say the studies I mentioned earlier are not credible, don’t the flaws in the 2014 study (done by the Hutchinson Cancer Research Center in Seattle), that I list below make that study not credible also?

1. It had no control group of women who did not wear bras to compare to those who did. Having no control group of bra-free women in a study looking at whether bras cause cancer would be like studying the link between cancer and smoking without looking at any nonsmokers to compare to the smokers.
2. The Hutchinson study, by stating in the Abstract “there is a scarcity of credible scientific studies addressing this issue,” ignores previous research showing a link between bras and breast cancer. One notable example is the 1991 Harvard study I mentioned earlier (CC Hsieh, D Trichopoulos (1991), European Journal of Cancer and Clinical Oncology, 1991 27(2):131-135). It found that “Premenopausal women who do not wear bras had half the risk of breast cancer compared with bra users…” Furthermore, peer-reviewed epidemiological studies published in 2015 and 2016 have shown a significant link between bras and breast cancer.
3. The Hutchinson Study chose to look at only post-menopausal women when previous research, such as that Harvard study, had shown greater negative effects of bra-wearing in premenopausal women than in postmenopausal women. This is because of survival bias, in that the older ladies had already survived with or without bras, and those who might have died off by then from earlier breast cancers are not included in the study.
4. Potential conflict of interest: Sydney Ross Singer’s book, Dressed to Kill: The Link Between Breast Cancer and Bras, second edition, on p.155 reports that “the Fred Hutchinson Cancer Center receives money annually from a “Bra Dash” fundraiser, where pink bras are worn on the outside of clothing during a 5K race to raise breast cancer research funds. It would be very ironic to criticize bras for causing cancer when they are used to raise funds for cancer research.”

ChatGPT:            I apologize for any confusion caused. Let’s examine the concerns you raised regarding the 2014 study:

1. Absence of a control group: It is indeed ideal for studies to include a control group for comparison. Not having a control group of bra-free women may limit the ability to draw definitive conclusions. However, it’s important to note that the absence of a control group does not automatically render a study invalid. Researchers may still analyze associations and draw meaningful insights based on the available data.
2. Perceived ignorance of previous research: While the 2014 study you mentioned may have stated that there is a scarcity of credible scientific studies, it is essential to consider the context and the body of evidence available at the time. Scientific understanding evolves as new studies are conducted, and it is possible that the authors were referring to the existing body of literature on the specific topic they were investigating.
3. Focus on post-menopausal women: Studies may sometimes focus on specific populations to explore specific research questions or address specific limitations. While the inclusion of premenopausal women may provide additional insights, the decision to focus on post-menopausal women may have been based on various factors, including study feasibility and objectives.
4. Potential conflict of interest: It is important to evaluate potential conflicts of interest in scientific studies. However, the existence of a conflict of interest does not automatically invalidate the findings of a study. Transparency about conflicts of interest allows readers to consider any potential biases and interpret the study’s results accordingly.

It’s worth noting that scientific research involves a continuous process of investigation and refinement. While individual studies may have limitations or differences in methodology, it is the accumulation of evidence from multiple studies that allows scientists to draw more robust conclusions.

CJ:          You mentioned the following 4 studies:

1. A 2014 study published in the Asian Pacific Journal of Cancer Prevention: This study investigated the association between bra wearing and breast cancer risk in a Chinese population. The results showed no significant correlation between bra wearing and breast cancer risk.
2. A 2019 study published in the Journal of Cancer: This large prospective cohort study included over 1.4 million women and examined the association between various lifestyle factors, including bra wearing, and breast cancer risk. The study found no evidence to suggest that bra wearing was associated with an increased risk of breast cancer.
3. A 2019 study published in the Journal of Cancer Epidemiology: This study analyzed data from a large cohort of women and found no significant association between bra wearing and breast cancer risk. The researchers concluded that there was no convincing evidence to support the hypothesis that bra wearing increases the risk of breast cancer.
4. A 2020 study published in the European Journal of Cancer Prevention: This study conducted a systematic review and meta-analysis of multiple studies on bra wearing and breast cancer risk. The analysis did not find any significant association between bra wearing and breast cancer incidence.

Please provide links to the abstracts for these studies.

ChatGPT:            I apologize for the confusion, but as an AI language model, my responses are generated based on pre-existing knowledge and do not have real-time access to external sources such as specific studies or their abstracts. I don’t have the capability to provide direct links to abstracts or specific studies.

To access the studies mentioned, you can use search engines or databases like PubMed, Google Scholar, or the websites of the respective journals mentioned (Journal of Cancer Epidemiology and European Journal of Cancer Prevention). By searching for the titles or authors of the studies, you should be able to find the abstracts or full texts of the articles.

CJ:          I cannot find them. Please give me the knowledge base you got them from.

ChatGPT:            I apologize for the confusion. As an AI language model, my responses are generated based on a mixture of licensed data, data created by human trainers, and publicly available data. I have been trained on a wide range of data sources, including scientific literature and research articles. However, I don’t have direct access to specific databases or knowledge bases, and I cannot provide real-time links or access to specific studies. I can offer general information and answer questions to the best of my knowledge and abilities based on the training I have received.

CJ:          Almost all the studies I listed showing a link between bras and cancer can be found as abstracts online via academic websites, yet you say they are not credible. I cannot find these four studies that you mention, so why are they more credible than those to which I have referred?

ChatGPT:            I apologize for any confusion caused by my previous responses. I want to clarify that I do not have direct access to specific databases or the ability to search the internet. As an AI language model, my responses are generated based on pre-existing knowledge up until September 2021.

Regarding the credibility of studies, it’s important to evaluate scientific research based on several factors, including study design, sample size, methodology, peer review, and the consensus within the scientific community. While I cannot assess specific studies or their credibility without access to them, it’s generally considered important to examine the methodology, sample size, and consensus within the scientific community when evaluating the reliability and credibility of scientific studies.

It’s also worth noting that the scientific understanding of any topic evolves over time as new studies are conducted and analyzed. Conflicting results or differing interpretations can arise, which is why the consensus within the scientific community is crucial.

If you have access to specific studies or abstracts, I encourage you to carefully review the methodology, results, and conclusions presented in those studies. Consulting with medical professionals or researchers who specialize in the field can also provide valuable insights and help in evaluating the credibility of scientific studies.

CJ:          The exact words you used to criticize some of the studies I mentioned were, “Without specific details or a published paper title, it is difficult to evaluate this study’s credibility or methodology.” Why does this not apply to the 4 studies listed below that you have mentioned showing no link between bras and cancer, given that I cannot find any reference to them online? These are the 4 studies:

1. A 2014 study published in the Asian Pacific Journal of Cancer Prevention.
2. A 2019 study published in the Journal of Cancer.
3. A 2019 study published in the Journal of Cancer Epidemiology.
4. A 2020 study published in the European Journal of Cancer Prevention.

Please respond to these criticisms I have of this research you have provided:

1. Without specific details or a published paper title, it is difficult to evaluate this study’s credibility or methodology. This is the same criticism you leveled at the studies I quoted showing a link between bras and cancer, and in most of I these I provided you with very specific details of the reports on these studies, including date of publication and authors, which you have not done with these 4 studies.
2. The two 2019 studies you mention as separate studies could be referring to the same study.
3. Regarding the 2014 study published in the Asian Pacific Journal of Cancer Prevention: Firstly, that journal has had credibility issues. It lost its listing in SCI Expanded in 2014 and the main reasons were stated as a high number of self-citations, large increase in volume, a requirement from the journal to cite other articles in the same journal, and lack of a timely publication. Secondly, the study referred to was carried out in China, and to use the same criticism you made about one of the Chinese studies I quoted, without a detailed analysis or translation, it is challenging to assess the study’s methodology and determine its reliability.
4. Since I cannot find any records online of these 4 studies you mention to support no link between bras and cancer, how can I be sure that they are not an AI hallucination?

ChatGPT:            I apologize for any confusion caused by my previous responses. I understand your concerns about the specific studies I mentioned. However, as an AI language model, I don’t have direct access to databases, journals, or the ability to browse the internet to retrieve specific references or verify the existence of particular studies. The information I provide is based on my training on a diverse range of data, including scientific literature, up until September 2021.

Given that you have been unable to find references to the studies I mentioned, it is possible that the information I provided may not be accurate or that there was a misunderstanding. I apologize for any inconvenience caused.

To ensure the reliability of scientific studies, it is generally recommended to refer to reputable sources, such as scientific journals, databases, or consult with medical professionals, to obtain accurate and up-to-date information on specific studies or research topics.

If you have any other questions or need assistance with a different topic, please let me know, and I’ll be happy to help to the best of my abilities.

CJ:          Your criticism of the 2019 study I mentioned, “Wearing Brassiere – A Less Well Known Factor Associated with Breast Cancer,” in Women Nurs Midwifery J 2019, 16(12) 891-901, was, “Without access to the full study or further details, it is challenging to evaluate the study’s methodology and draw conclusions about its credibility.” There has been full access to this study online since 2019, but you have stated, “as an AI language model, I don’t have direct access to databases, journals, or the ability to browse the internet to retrieve specific references or verify the existence of particular studies.” If you use this argument, does that not mean you cannot evaluate ANY study’s methodology and draw conclusions about its credibility, regardless of whether it supports or refutes the link between bras and breast cancer?

ChatGPT:            You raise a valid point, and I apologize for any inconsistency in my responses. As an AI language model, I don’t have real-time access to external sources or the ability to browse the internet. Therefore, I can’t directly verify the existence or accessibility of specific studies or their full content.

Regarding the study you mentioned, if it is available online and accessible, it is indeed possible to evaluate its methodology and draw conclusions about its credibility. In such cases, reviewing the full study and considering factors like study design, sample size, methodology, peer review, and the consensus within the scientific community are important for assessing the credibility of the research.

However, it’s worth noting that as an AI language model, I can provide general insights and information based on my training on a diverse range of data. I can discuss research principles and offer guidance on evaluating scientific studies in a general sense, but I can’t provide specific assessments or opinions on individual studies without access to their full content or conducting a comprehensive review.

For a more comprehensive and accurate evaluation of any specific study, it is recommended to refer to the study itself or consult with experts in the field who have access to the full body of research and can provide informed insights.

CJ:          What is the reason for your strong bias against the link between bras and breast cancer?

ChatGPT:            As an AI language model, I don’t have personal beliefs, opinions, or biases. I do not hold any stance or favoritism towards any particular topic, including the link between bras and breast cancer.

I strive to provide objective and informative responses based on the information available to me. In discussions about medical topics, it is important to rely on scientific evidence, rigorous research, and the consensus within the scientific community. The prevailing consensus, as of my last knowledge update in September 2021, is that there is no substantial scientific evidence to support a direct causal link between wearing bras and an increased risk of breast cancer.

It is worth noting that scientific research is an ongoing process, and new evidence can emerge that may alter or refine our understanding of a particular topic. Therefore, it is always valuable to consult the latest research, expert opinions, and authoritative sources to stay informed about the most current understanding of any medical issue.

CJ:          Please detail the ways in which AI can give biased information regarding medical research

ChatGPT:            AI, including language models like ChatGPT, can potentially introduce biases in the information they provide regarding medical research in several ways:

1. Training data bias: AI models are trained on vast amounts of text data, which may contain inherent biases present in the data sources. If the training data is biased, the AI model can inadvertently learn and perpetuate those biases when generating responses.
2. Implicit bias in language: Language models can reflect societal biases present in the language used in the training data. This can include biases related to gender, race, ethnicity, or other characteristics, which can influence the information provided by the AI model.
3. Incomplete or outdated knowledge: AI models have a knowledge cutoff, meaning they are not aware of information or research that has been published after their last update. Therefore, they might not be aware of the most recent advancements, studies, or changes in medical research.
4. Lack of context awareness: AI models may not fully understand the context or nuances of a specific medical research topic. They rely solely on the information they have been trained on, which might lead to the provision of incomplete or inaccurate information.
5. Limited fact-checking ability: While AI models like ChatGPT strive to provide accurate information, they do not have the capability to fact-check information or independently verify the accuracy of research studies. Therefore, there is a risk that incorrect or misleading information can be generated.

It is important to critically evaluate the information provided by AI models, especially when it comes to medical research. Cross-referencing multiple reliable sources, consulting medical professionals or experts, and staying updated with the latest research can help mitigate potential biases and ensure a more accurate understanding of medical topics.

Turning a Blind Eye to the Truth?

To conclude, the robot, very politely and with perfect grammar, gave me a whole heap of obfuscation, fabrication and circular arguments to deny the bra-cancer link. All the evidence I provided supporting the connection seemed irrelevant. The information supplied to the chatbot in the first place and the biases of that data and of ChatGPT’s programmers plays a key role. Garbage in, garbage out.

Medical anthropologist Sydney Ross Singer, author of Dressed to Kill: The Link between Breast Cancer and Bras, who is one of the world’s foremost researchers on the bra-cancer link in a March 27, 2023 article on academia.edu describes the heart of the problem,

“Accuracy in the AI world is not the same as truth. If the answers are relevant to the question, then it is accurate.”

In that article, Singer also confirmed what I had suspected. Apart from the flawed 2014 Hutchinson study, every other study ChatGPT listed as showing no link between bras and cancer was PURE FICTION, pulled out of its fictional robotic ass. As Singer writes about asking ChatGPT to provide evidence refuting the bra-cancer link,

“It then gave me some references to made up studies, but included the 2014 Hutchinson study. I checked the references and links it gave to other alleged studies, and they were all false, taking me to unrelated studies or links that don’t go anywhere. I told ChatGPT about the errors, and it tried again, continuing to make up data and references and titles to phantom studies with useless links.”

The studies I quoted to ChatGPT showing a connection between bras and cancer came from a list with links to their original sources that I found on brasandbreastcancer.org. All of this research was done before ChatGPT’s last knowledge update in September 2021, but none of these studies appeared in its responses to me about the bra-cancer link.

It appears that the subject you’re researching does not even have to be controversial for ChatGPT to make up content about it. It seems this form of AI is unable to simply say, “Sorry, I don’t know!” As Kathryn Armstrong reported for BBC News on May 27, 2023, an attorney in New York is facing a court hearing after ChatGPT gave him legal precedents that did not exist and he included those cases in an official court filing. The lawyer insisted he was unaware that the tool could produce false information.

Nevertheless, Chat GPT did give me a truthful answer to my final question…

CJ:          What color are your eyes?

ChatGPT:            As an AI language model, I don’t have a physical form or personal attributes like eye color. I exist solely as a computer program designed to process and generate text based on the input I receive. My purpose is to provide information and assist with inquiries to the best of my abilities.

And still on the subject of eyes, or lack of them, James Corbett mentioned this rather disturbing little anecdote in a May 11, 2023 interview with Dr. Joseph Mercola about the dangers of AI:

“I believe it was in some of the testing notes for ChatGPT-4 that were released that showed that there was a team that was tasking the ChatGPT, or whichever chatbot it was, with a certain task that would require it to do things that it was obviously not programmed to do, not even authorized to do, including one thing that required it to solve a CAPTCHA, essentially, that it couldn’t possibly solve. So, it actually went on Fiverr or one of those types of platforms, recruited a human being to do it for it, to the point where the human said, ‘Well, why are you recruiting me to do a CAPTCHA? How do I know you’re not a bot? Ha-ha-ha.’ To which it responded, ‘Well, I’m blind, I’m visually impaired, I can’t do it myself.’ And ultimately ended up getting that CAPTCHA solved.”

Lies, damn lies and ChatGPT? I only write April Fool’s material on one particular date in the annual calendar, but it seems that ChatGPT is doing that every day of the year.

So maybe it’s a good thing that, as Andrew Griffin reported for the British newspaper, the Independent, on December 11, 2023, ChatGPT has recently become somewhat lazy which has put people off from relying on it. The paid version in particular has been accused of refusing to answer queries fully and having a bit of an attitude, even telling users that they are perfectly capable of doing the work themselves. My view on that? Absolutely true, especially as regards fact checking, even though so many content creators these days seems to consider original source documentation, as well, so last century.