We’re planning a trip. I asked ChatGPT when we should buy tickets. ChatGPT gave answers that seemed based on data. I could have found out the answer using a trip planner like I usually do, but ChatGPT might save time without leading me seriously astray.
I asked ChatGPT about treating separation anxiety in my dog. ChatGPT presented things that are science-based AND things that are not science-based.
Trying to follow ChatGPT’s many recommendations would be exhausting for the dog parent and ineffective. The training plan ChatGPT provided would have made separation anxiety worse for our dog by pushing duration too fast for my dog to feel safe.
In my last job, I used machine learning and AI. I learned how to tell when a model can be relied on and, if not, why not. ChatGPT and other AI bots are just not helpful when you need an answer that is supported by science, matches your values, and is customized for you and your dog right now.
Why do I not recommend ChatGPT or other Generative AI for dog training advice?
ChatGPT cannot tailor advice to you and your dog, in this moment, with skill. Even when I provided lots of individual info in my prompts, what ChatGPT offered was not skilful.
ChatGPT is not discerning. ChatGPT takes in everything it can find and makes answers that are kinda like that everything. That often includes making some stuff up – not a big deal when I want tips on searching for flights, yes a big deal when I have a dog who is chewing the door with panic.
Unfortunately, when it comes to dog training, there is a LOT of misinformation and opinions stated as fact.
By pure volume, there is probably more misleading info out in the world about dog training than there is accurate info. That’s a problem for ChatGPT.
Dog training includes, on one side, people who use pain and fear to compel dogs to do something.
If you’ve used pain and fear to motivate your dog, you’re not alone. That’s what I learned with my childhood dog. We do the best we can with the information we have.
Pain and fear work great for suppressing behavior. Unfortunately, it feels super icky to use these methods – as I remember from childhood. And prong, shock, and the like come with side effects: increased fear, anxiety, and aggression.
On the other side is trainers and dog parents who use things dogs like, often snacks, to motivate the dog. Force-free, reward-based trainers like me do NOT use any fear or pain. Positive reinforcement training like ours is highly effective, strengthens your connection with your pup, AND comes without any harmful side effects.
ChatGPT does not know how to distinguish between these two ways of training dogs. ChatGPT does not have values like kindness and compassion to decide which ways of training are in and which are out.
I want to be really clear that force-free trainers like me do not have a monopoly on the skills to train your dog. One of my main goals, in these posts and if we work together, is to support you in building those skills. And to know that you are a good dog parent.
If you have a question about dog behavior or training, and you’d like an answer from a human, ask me! I’ll reply and may write a post about it.
