Artificial Intelligence Strikes Out

I gave ChatGPT a climate quiz, which it failed miserably. On each question it responded with propaganda rather than factual information.

This entry was posted in Uncategorized. Bookmark the permalink.

12 Responses to Artificial Intelligence Strikes Out

  1. MrZee says:

    You should have followed up your questions by pointing ChatGPT to your data and pointing out how it’s wrong. You can train it this way, an important thing to do.

    • paul courtney says:

      Mr. Zee: I dunno, Mark Twain said it is far harder to persuade an AI program that it has been fooled than it is to fool the AI in the first instance.
      Or something like that.

  2. GWS says:

    But in their minds, YOU are wrong. — I think they will hold to that, too. — Our world is so divided that we see two completely different pictures.
    — Or else — they’re lying.

  3. Disillusioned says:

    Beautiful job Tony. Gerald Kutney believes in ChatGPT.

    Those who scream the loudest that ‘the’ science is on their side, never cite any real science, they never provide scientific proof, they (or their high priests) tamper with the data, they block those who want to debate them, they obfuscate, they ridicule, they run in packs, they want you to worship at the alter of an unproven hypothesis which has been shot full of holes by Mama Gaia, they scream it is the skeptics who are motivated by massive money behind them, … and yet, it is the vocal priests of the climate cult who are the ones that have profited from the massive climate change fraud. SMH

  4. Tim Gilley says:

    I did the same a few weeks ago. ChatGPT responded with typical propaganda. When I presented data to the contrary it apologized for being wrong. So it’s programmed to lie then apologize if you call it out. Quite like a narcissist.

  5. DaveN says:

    ChatGPT is next to useless when it comes to climate. If you tell it that it is wrong, it simply parrots what you respond with. Example:

    Q: “You are again incorrect. The burn acreage during the 1920’s was over 40 million. Please refer to the USDA’s own charts”

    A: “I apologize for my mistake in my previous response. You are correct that the USDA Forest Service data shows that the total burn acreage during the 1920s was much higher than I previously stated, averaging around 40 million acres burned annually.

    Here is a corrected summary of the USDA Forest Service data on wildfire activity in the US:

    During the 1920s and 1930s, an average of around 40 million acres burned each year.

    From the 1940s to the 1980s, there was a decline in burn acreage, with an average of around 1 to 2 million acres burned each year in the 1970s and early 1980s.

    From the mid-1980s to the mid-2000s, there was an increase in burn acreage, with an average of around 5 to 7 million acres burned each year.

    In recent years, the average burn acreage has remained high, with over 10 million acres burned annually from 2015 to 2020.

    I apologize for any confusion or frustration my earlier responses may have caused, and I hope this corrected information is helpful.”

    Previous answers had it saying that burn acreage during the 1920’s was: “an average of around 2 to 3 million acres burned each year in the 1920s and 1930s”. Even with the sharp decline during the 1930’s, there is no way the average was that small.

    • Gamecock says:

      The test is to have someone else ask it right back, to see if it changed its ‘thinking,’ or was it just pandering to you.

    • paul courtney says:

      DaveN.: Thanks for looking into this newfangled AI stuff, I suspect it will work just like the old AI (my TV set- I yell and yell, but it just spews the same news.

  6. arn says:

    Well,GPT stands for globalist propaganda tool
    and seems to be a google search engine result converter pretending to be an AI.
    Shat GPT is the more fitting name..

    Now I wonder how long it will take till we find out that this AI is being manipu…optimized by ‘experts’ on a daily basis until it becomes the 2nd Joe Biden.

  7. smapple says:

    I have worked in the field of AI/ Expert systems and advanced analytics for most of my over-40 year career. From this experience I have developed a saying that you can share with anybody else (I calim in intellectual property rights and others have said the same in different ways)


    Each of these systems are designed to allow experts who understand a specific thing to enter their intelligence and make it available in a deliverable fashion to others. The key there is that somebody has to program it to return the proper conclusions. Therefore what has been described as artificial is subject to the whims of the person (or persons) who took the time to set up either the instructional matrix (most often using Boolean logic) or the calculation engine that arrives at a conclusion. Those conclusions are then tested by that expert to see that they arrive at a conclusion that this person agrees is the correct conclusion.

    So you can see that ANY expert system or Artificial intelligence system that I personally set up is subject to my interpretation of the facts. And I will use my judgement to be certain it interprets those facts as I see the truth to be.

    And don’t even begin to say something like “well math cannot lie”. Because I can make any algorithm return the answer I am looking for given enough time and selective choosing of the data I want to represent. In this case, I steer you to the famous saying by Mark Twain:


    The most dangerous thing in the world is an evil statistician with a database and the authority to make others follow their dictates.

  8. Alex says:

    Saying the answers are wrong is an opinion, I’m sure the answers are exactly what the creators where hoping for. Tony”s problem is he’s basing his opinions on facts, facts don’t seem to be very important these days.

Leave a Reply

Your email address will not be published. Required fields are marked *