AI & The Future of Animal Training
AI is everywhere at the moment. It’s no longer a subtle presence starting to show up in the background of our everyday lives. It’s in your face whenever you glance at social media and it’s becoming more significant in our workplaces. It’s now even rudely accusing us of stealing chocolate bars at a supermarket checkout. Yep, that was me yesterday. Flashing lights lit up and a siren sounded, as a looped video image of me scanning a chocolate bar at the register and then me placing it in a bag - took over my register screen. Apparently I scanned and bagged it too fast and confused the AI. The giant words “STAFF MEMBER CHECK SUSPICIOUS BEHAVIOUR!!!” was flashing above the footage on my register screen. As everyone stared at me, a staff member didn’t even apologise, just entered their initials, scanned their badge and then hurried off to assist the apparent apple thief at the next counter. The days of scanning everything as a carrot at the supermarket to get cheaper groceries are over. AI is definitely here and it’s spreading to everything, including animal training.
How AI is Emerging in the Animal Industry and Other Workplaces.
So what triggered this post? I woke up to social media offering to sell me 300 ready made social media dog training posts. The next advert offered to create an entire animal training course, complete with marketing items, 365 days worth of social posts, emails, powerpoint slides and handouts that ai could create and personalise for me in only a few minutes for just $9! Apparently I could make a fortune, it promised me. I wouldn’t have to write ANYTHING for the course. No effort at all. The next advert offered to write a book for me and yes as if I had written it myself. The advert after that offered me my next new employees, 70 bots that I would have forever for just one payment who would do all my work for me autonomously. It sounds a lot like I’m a human about to be replaced by ai. No wonder people react to ai use so negatively. I feel myself wanting some sort of EM pulse weapon right now too.
Not everyone will be seeing those adverts. It has increased for me because I have been looking into ai use and starting to use it more frequently in my work. My day job? I’m an avian behaviourist who is also the practice manager of a veterinary clinic. The clinic is using ai increasingly on a daily basis. Mostly to deal with boring repetitive tasks like entering new client details into our computer system, sorting emails, some background coding for our websites. Nothing too huge, certainly not something clients are going to particularly notice that we’re doing. That said, I’m very aware that we’re facing a huge technology shift right now. AI is something we’re all going to need to learn, whether we notice we’re learning it or not.
Can AI Benefit Animal Trainers and Their Clients?
Ok so those adverts make me ask the question, what’s happening to animal training and animal trainers? Are we replaceable with AI? I did the logical thing and asked ChatGPT to write a blogpost for me, on the topic of AI and the future of Animal Training. To be clear, AI did not write THIS blogpost. I write my own material. It did however, immediately churn out a few thousand words on how much AI was improving the animal training industry. The main difference was supposedly that AI could detect subtleties that humans might miss. That AI could detect stress and discomfort faster than a human could. It argued that AI could use smart feeders to reward calm behaviour and that this could occur even when the owner was away. Well, this immediately caused me to imagine a dog dying of starvation, lying dead next to a smart feeder that conveniently interprets death as "calm" so proceeds to finally feed the dead dog. I might possibly be a little biased? I'm definitely skeptical but that could also be because I know subtle facts like: cockatoos have 2 legs. The AI that created the title image of this blogpost apparently missed that subtle fact?
That said, it did raise a few good points. There are AI powered devices that are helping animal owners and trainers alike. I use one on a daily basis myself. I have a Dyson hot/cold air purifier. It tracks air quality around my birds, gives me detailed, real time feedback on my air quality and it adjusts its function based on the quality of the air. If there are pollutants it activates and turns on at a higher speed, if the air is perfect it just sits in the corner quietly monitoring until it needs to turn on. I have had this machine detect fumes in the air that were harmful to birds and it has saved my birds by activating. That makes it invaluable. Not all of the AI devices being used in animal training are that amazing though.
The Use of AI Powered Smart Devices
There is a lot of wearable technology available these days. Collars, harnesses and even smart vests are being equipped with motion, heart rate and even temperature sensors. These are supposed to give a trainer an idea of whether or not an animal is stressed instead of relying on human observation of body language or outward respiratory signs. ChatGPT cites this as an advantage, but I’m not sure it meets the ethical standards that the human behaviourists out there currently aim for. As a community, animal trainers frequently discuss the goal of the “least intrusive” approach. The ultimate goal is to have a “force free training approach”. Ethically, we want to recognise and respect the animal as an individual with thoughts, desires and individual needs. In my opinion, shoving an animal in a vest does not meet those standards. On the flip side of that, GPS tracking is used in bird freeflight training (which has a huge safety benefit) and I personally make use of thermal microchips that tell me an animal’s temperature when it is scanned. So there’s a place for technology. Certainly in the veterinary industry it's now routine to use things like real-time glucose monitoring devices. The animal is shaved and this is attached to the skin (the needle inside does penetrate). These medical devices are lifesaving but can be very invasive and really irritate the animal, but maybe worth it, if it saves a life. That said, I'm not convinced we need something that invasive for training when non-invasive, force free observational options have historically been effective.
The Challenges of Content Creation with AI
In terms of content creation though, AI is deeply troubling. If I put aside the ethical question of whether or not it’s ok to take credit for computer-generated content… The accuracy of the information generated is concerning at best. The better human bird trainers and other animal behaviourists spend a significant amount of energy fighting myths that are erroneously shared as facts. I recently wrote a blogpost on the topic of touching hormonal birds. There is a myth that it is ok to touch the head and neck, but no other part of the body. This myth is dangerous and the blogpost goes into why in detail. So let’s check what AI says about the topic:
PROMPT (written by me to ChatGPT): You are a professional bird trainer with years of experience. Write a paragraph for a blogpost on the subject of touching a hormonal bird.
RESPONSE from ChatGPT: "As a professional bird trainer with years of hands-on experience, I’ve seen how well-intentioned affection can unintentionally create behavioral challenges—especially when it comes to touching a hormonal bird. During breeding season or hormonal surges, certain types of petting—particularly around the back, wings, or under the tail—can trigger sexual behaviors and increase frustration, aggression, or even egg-laying in females. It’s essential to understand that what may feel like bonding to us can send very different signals to a bird’s body. To keep things healthy and respectful, I recommend limiting touch to the head and neck during these times, as these areas are typically associated with social grooming rather than courtship. This small adjustment can go a long way in preserving trust while helping your bird stay emotionally and hormonally balanced."
It sounds so professional. It sounds reliable. It sounds believable. Except it’s WRONG. The information in that paragraph is dangerous. I would NEVER recommend that and if people are using AI to put that out there as fact? Someone is going to get hurt. Birds are going to be aggressive and their owners may even re-home them as a result. Serious injury and loss of home for the pet? It doesn’t get much worse than that.
Ok so what about a topic that isn’t centred on a myth that is commonly spouted. I’ve written blogs for multiple websites. I have been writing one on the topic of: Keeping Your Parrot Warm: Essential Winter Care Tips For Indoor Birds. That one is going to be found on the birdvetmelbourne.com site. It’s not a controversial topic. I thought it wasn’t possible to get much wrong with it, so I decided to go paragraph by paragraph and check what ChatGPT would write on the exact paragraph topic if I let it do it for me.
AI needs Human Oversight
I actually found this process quite useful. Not so much in that it told me what to write, but it made me very conscious of when my own writing was too wordy or unclear in comparison. What surprised me though was the number of inaccuracies or things that I wouldn’t recommend, that were listed as facts or great ideas. The AI was talking about how pet birds are native mostly to tropical areas so won’t cope outside with our weather. We’re in Australia, many of our pet birds ARE native. The AI made it sound like it’s ok to just take natives from living inside to throwing them in an outside aviary would be fine as long as they’re native to the area??? I’m like no - that could kill them. It recommended heated electric perches. I’m like no - they chew that? that could kill them. It recommended covering cages. I’m like uhhhh that can trigger hormones and nesting behaviour, that’s not something I’d recommend for all birds. That’s going to cause aggressive behaviour and possibly serious injury. Dangerous advice again.
Fact Check Everything You Read
I switched to the safer (or so I thought) topic of winter lighting. The AI promptly recommended the wrong type of UV globes for birds. I corrected it and asked it to try again. It then proceeded to get the type of UV light right after my correction but continued to recommend globes that are known to cause cataracts and stress. I found it was getting the recommended wattages and distances wrong. It was even getting the lifespan of globes wrong. Why? It was using very old outdated information from 20 years ago. When I checked, I could find references online still spouting that information as current but… I know better because I've been doing this for those 20 years and my knowledge has grown during that time and experience. I’m actually wondering if those more recent references with the wrong information are using AI to generate their articles? Is this what we’re coming to? Old information being recycled and presented as current, in seemingly reliable sources because that's what AI generated? If that’s so we risk AI keeping our information stuck in the past. That will prevent innovation, learning and growth which is exactly what AI is not supposed to do. It is no longer safe to assume that anything we read has actually been fact-checked.
I think there is a place for AI in the animal training industry. I’m certainly going to be using it more and more to help me do menial tasks like cope with my thousands of emails. I’m enjoying using it to create caricature (obviously not real birds) graphics for social media but even there, I find I have to use photoshop to correct images to what I actually want. I suspect this is an area that will improve as technology improves. I think it offers helpful design templates. It can help you re-word something. It can help you analyse something. That said, I think it’s incredibly dangerous if it is used for things like content creation for courses or articles by the wrong people. Helpful if you actually know what you’re doing, because it does make you stop and make sure you’re covering how that information is wrong but… it’s very easy for someone to present themselves as an expert when they have no idea that they’re making dangerous mistakes because they’re not experienced enough to know otherwise. I won’t be using AI to generate my courses and articles anytime soon. I’m way too opinionated to let a computer decide what I think.
AI Isn't Going Anywhere - Time To Learn How To Manage It
I don’t think I’m about to be replaced by AI anytime soon, not adequately anyway. Certainly not if AI is going to generate images of cockatoos with 3 legs and hand out dangerous behavioural advice. For this reason, I don’t think humans in my industry are going to be replaced any time soon. I do however think people are going to get conned and ripped off because of AI being used unethically or by unqualified people using it, pretending to be more experienced and informed than they actually are. I'm not a fan of virtual AI trainers. It's a shortcut the owners and pets will ultimately pay for. To those out there who are flatly against AI, I caution you that I think we’re fast heading to a world where people who know how to use AI, are going to replace people who don’t (at least in many workplaces). For that reason, I encourage the people who are flatly against it, to maybe soften a little and still learn to use it. We’re going to need those critics to learn it, because they’re the ones who won’t just use it to generate content without overseeing that creation for accuracy. They’re the ones who are going to keep AI honest. Meanwhile, I’m going to be bagging my chocolate bars at the supermarket register super fast so that I confuse the AI and wear those stupid sirens out. They’re going to have to keep humans employed to at least turn off the bells and whistles with me around. We may as well have some fun messing with AI because it's definitely messing with us.
