Artificial Intelligent (AI) programs have been all the rage for a couple of years now and they are being used more and more by gardeners to get their information. They are easy to use. Just enter a question and an answer appears. That is a lot easier than doing a Google search and having to read through blogs to get the information, but is it a better way? And most importantly, is the information correct.
Here are Garden Myths we don’t accept every answer, even if it is from AI, so we decided to do some testing.
NOTE: Since I published this post – some people have been getting errors saying he site is not safe and may have a virus – I have checked it several ways and it doesn’t have one. In an effort to try and resolve the problem, I have removed all the links in this post.
Does AI Provide Reliable Information?
One subject that I know very well is the fact that plants DO NOT clean the air in our homes. We asked ChatGPT, a popular AI application, this question, “Do plants purify the air in your home?” This is what AI had to say:
“Yes plants have the ability to purify the air in your home to some extent. ….
Much of the information in the reply is correct. Plants can absorb pollutants, they carry out photosynthesis, they produce oxygen, effectiveness depends on size and number of plants, etc. However there are two glaring errors.
- Plants do not purify the air in a home – see my blog post about air purifying plants.
- Plants do not increase the oxygen level in our homes – see my blog post about oxygen levels in a home.
AI Makes Excuses for Its Mistakes
What happens if you challenge AI? To find out we asked another question. “But I read, “A reasonable amount of plants in the home or office will NOT purify the air.”” The quote is taken right from my blog post – this should allow AI to find my post and read it.
ChatGPT says, ” As an AI language model, I do not have access to current research or the ability to verify specific statements …” That is a huge red flag. It admits it doesn’t really know what the right answer is.
It goes on to say “it is generally accepted among scientists and researchers that …..”. Wait a minute! It just said it does not have access to research and that is where most ideas from researchers would be located. It just contradicted itself.
It then says, “It is always recommended to consult credible sources ….” So ChatGPT does NOT consider itself to be a reliable source. After getting your answer you need to look somewhere else for the truth!
AI Mumbo Jumbo
Avery Rowe on Tradescantia Hub did some testing of the AI called GROOT. He asked, “Which plants will decrease humidity in my house?”, and got this as an answer:
“To decrease the humidity in your house, you should look for plants that have high transpiration rates. Transpiration is the process by which plants release water vapor into the air. Plants with high transpiration rates will release more water vapor, which will help to decrease the humidity levels in your home.”
GROOT clearly understands transpiration and how plants use it, but fails completely in making the logical connection between transpiration and humidity. Transpiration increases humidity!
Avery Rowe also asked GROOT to describe the flower structure of Tradescantia zebrina and provide scientific sources. GROOT got the color of the flower wrong, but worse than that it listed three references to support its answer. Problem is that none of the references actually exist. AI just made them up.
Aaron, a member of our Facebook Group, Garden Fundaments, asked ChatGPT for some information about squash vine borers and nematodes, and asked for a reference. He was given one: “Clausen, L., Koppenhöfer, A. M., & Noling, J. W. (2014). Efficacy of entomopathogenic nematodes for control of squash vine borer, Melittia cucurbitae (Lepidoptera: Sesiidae), in greenhouse experiments. Journal of Nematology, 46(1), 1-7.” Neither Google nor Google Scholar can find such a reference.
This seems to be a common occurrence with AI programs – they make up all kinds of facts, including dates of events, names of people, and references. In one case it made up a newspaper article in The Guardian to support its lies, complete with publication date, authors and quotes.
What Does AI Think of Itself?
A new AI app for garden answers was released by Garden Savvy, so I asked, “Is your information always correct?” I know it is not correct because I asked it about plants cleaning the air in our home and it gave about the same wrong answer as ChatGPT and I asked it about marigolds as companion plants and it got that wrong too. Here is what it thinks of itself:
“I strive to provide accurate and reliable information based on the knowledge and data programmed into me. However, it’s important to note that I do not have real-time access to the internet or the ability to verify the accuracy of every piece of information. My responses are generated based on pre-existing knowledge and patterns, which may not always reflect the most up-to-date information or account for every possible scenario. Therefore, while I aim to provide accurate information to the best of my abilities, it’s always a good idea to cross-reference and verify the information I provide with authoritative sources or consult experts in the respective field to ensure the most accurate and current information.”
Some key points:
- It does not have access to the internet.
- It can’t verify information.
- You have to verify the information with experts – AI is not an expert.
At least it seems to be honest about itself.
What Does AI Think of Me?
Gary from our Facebook Group, Garden Fundaments, asked AI about the video below, and then asked it, “Which garden myths do you agree with?” It replied with:
“I do not agree with any of these myths because they are not supported by science. … I do respect Robert Pavlis for his efforts to educate gardeners and debunk myths with logic and research.”
This proves AI is quite smart!!!
The Problem With AI
The AI technology I’ve discussed is known as “generative AI”, or “Chatbots” and they are designed to take words and string them together into a proper sentence that seems to be correct. That is why its answers sound so authoritative and believable.
AI is not designed to think through a problem and solve it. It is not designed to take facts and bring them together into a logical scenario. It is designed to find related words and string them together into a good story.
You might think that AI has access to all of the knowledge on the internet, but that is not true. It has been trained on a very select amount of information and the cutoff date for that information was October 2021. In February 2023, the Harvard Business Review put it this way, “it doesn’t “know” anything beyond that. As far as ChatGPT is concerned, Russia hasn’t invaded Ukraine, FTX is a successful crypto exchange, Queen Elizabeth is alive, and Covid hasn’t reached the Omicron stage.”
From the above examples, and many more on the internet, it is clear that even the facts that AI presents may be completely wrong and fabricated. That is really scary to me. A lot of garden information is wrong because well meaning people simply can’t think through a problem or don’t understand the scientific method, but at least they reach the wrong conclusion with honesty. Honesty seems to be lacking in AI.
My fear is that AI will generate a lot of misinformation and we know that there are lots of gullible people ready to believe its nonsense. They in turn will repeat the information and present it as fact. That will lead to even more gardening myths. The only upside is that I will never run out of stupid gardening myths that need to be debunked.
It is worthwhile asking AI questions about things you understand well. You very quickly realize it has no clue what it is saying. However, the answers are pretty smooth and if you don’t know the subject matter you will be easily fooled into believing the answers.
Should you use AI? Use it to get answers and to get new words about a topic. For example, if you asked the above question about squash vine borers and nematodes, you would have learned that the nematode which attacks the squash borer is called Steinernema. You can then use this word to do some proper research to find the real facts. Don’t believe AI or repeat what it says unless you verify the information, and it goes without saying that for gardening information there is no better place than this blog 🙂