ChatGPT makes shit up

Warning! Chat GPT makes up false references and abstracts.

Recently I saw a post by a friend on social media about high levels of Bovine Leukemia Virus (BLV) in New Zealand cattle. He was concerned because BLV has been associated with breast cancer in humans. I was interested as I hadn’t heard of this, and asked for references, he then sent me two very legit sounding NZ journal references (details below). I looked up the articles to get details and found they did not exist. I got back to my friend, and he told me it was from Chat GPT. I did some more research and BLV has only ever been found only in a tiny proportion of NZ cattle

Let this be a warning as I have noted before not only is Chat GPT not able to discriminate between promotion/spin and reality, but I was shocked to find it just makes shit up!

He asked Chat GPT for a NZ journal article showing BLV in NZ (answer in italics)

Here's a reference to a New Zealand journal article that demonstrates the prevalence of Bovine Leukemia Virus (BLV) in New Zealand:

Title: Prevalence of Bovine Leukemia Virus Infection in New Zealand Dairy Cattle as Determined by a High-Throughput Enzyme-Linked Immunosorbent Assay

Authors: Ridler AL, et al. Journal: Journal of Veterinary Diagnostic Investigation. Year: 2015

Volume: 27, Issue: 2, Pages: 182-187

In this article, the authors investigated the prevalence of BLV infection in New Zealand dairy cattle using a high-throughput enzyme-linked immunosorbent assay (ELISA). The results showed that BLV was prevalent in New Zealand dairy cattle, with an estimated overall herd seroprevalence of 71.8%.

I looked up the paper and it does not exist; the authors and the abstract are completely made-up.

I told my friend it was fake, and he then asked – Chat GPT this:

Please check this journal reference. Is it genuine

- Paine, S., et al. (2020). The prevalence of bovine leukaemia virus infection in New Zealand dairy cattle. New Zealand Veterinary Journal, 68(2), 90-96.

Yes, the reference you provided is genuine. The study titled "The prevalence of bovine leukaemia virus infection in New Zealand dairy cattle" was published in the New Zealand Veterinary Journal in 2020 and authored by Paine, S. et al. The study investigated the prevalence of bovine leukemia virus (BLV) infection in dairy cattle in New Zealand and found that the overall prevalence of BLV infection was 58.6%.

So, it just doubled down on the lie. Then he asked Chat GPT4 this:

 “The best journal reference to show BLV is prevalent in New Zealand dairy cows”

One of the best journal references to show that BLV (Bovine Leukemia Virus) is prevalent in New Zealand dairy cows is:

Stewart, M., Laven, R., & Nicol, A. (2018). Bovine leukemia virus infection in New Zealand dairy cattle: risk factors and prevalence. New Zealand veterinary journal, 66(2), 86-92.

This study conducted a cross-sectional survey of 278 New Zealand dairy herds to determine the prevalence of BLV infection and identify potential risk factors. Blood samples were collected from 2,378 cows and tested for BLV antibodies using an ELISA test. The study found that BLV was present in 71% of the herds and 44% of the cows. The prevalence of BLV infection varied significantly among regions, with the highest prevalence in the North Island. The study also identified several risk factors for BLV infection, including herd size, age of cows, and the presence of shared grazing areas. Overall, this study provides strong evidence that BLV is prevalent in New Zealand dairy cows.

Once again I went to the journal and searched and this is a totally fake reference both authors and abstract were just made-up, I searched the journal could not find these authors. I searched for BLV in NZ herds and found a few old papers and it was found only in a tiny proportion of NZ cattle.