How AI Pet Videos Are Affecting Our IRL Dogs and Cats

When OpenAI released Sora 2, it heralded a new age of AI animal videos on TikTok and Instagram. The clips show everything from the popular ginger cat ‘Chubby’ to podcasting pet presenters, to small babies riding dogs around the house and doorbell cam ‘footage’ of trampolining wild animals. But so convincing is some of this AI-generated pet content, (OK, maybe not all of it is convincing) that it can confuse our grasp of what’s real and what isn’t. And those blurred lines could lead to real life consequences – for you and your pet.

Easy content creation

No stranger to AI content tools is Samantha Henry of online retailer Pup Chic Boutique, who says that she and many of her customers use them. Lots of the videos are created using CapCut; the user-friendly video editor owned by ByteDance, the China-based parent company of TikTok. CapCut offers direct sharing to TikTok, making it a go-to tool.

“CapCut has AI templates so you can upload a photo of your dog, and the app will turn it into video. There is one that is trending of someone holding a microphone and asking your dog what it wants to be when it grows up. And there’s another one that merges photos of you and your dog to create a photo of you touching heads,” says Samantha.

As well as driving engagement because of their popularity, Samantha can see how enjoyable these videos are. “These videos provide a little bit of fun. Some of them, like the heads together photo, are used a lot by people who have lost a pet, it allows them to have a beautiful memory of their lost pet,” she says.

Samantha believes there’s an added advantage to people using AI to create videos of their pets. “A lot of people are dressing their pets up for social media posts, some dogs enjoy it, but many don’t and a lot of animal charities don’t agree with pets being a fashion accessory, so this is a great way for people to create that content without forcing their dog to do anything,” says Samantha.

The benefits of AI pet content

Pet content isn’t new of course. The first viral video of a cat (1984’s Keyboard Cat) made it to YouTube in 2007 and by 2014 there were about two million cat videos on YouTube that had been watched 25 billion times. Data from 2023 shows the #cat hashtag on TikTok surpassed 500 billion views. 

Nina Fotara, a trainer at Confident Canine in Essex, knows that while some dogs (including her own) love to show off their abilities, not all animals like attention, and agrees that AI-generated content can reduce the inappropriate use of pets in social media videos, helping pets avoid the dangers posed by certain internet trends.

“This kind of pressure to perform can be stressful and, for the wrong dog, overwhelming,” she says, “Behaviourists and trainers can spot early signs of stress and advocate for the dog before things escalate, whereas those signals are often missed by well-meaning owners.”

“The main potential benefit of AI-generated pet content is the reduction of pressure on real animals, particularly where content would otherwise involve coercion, repetition or stress,” says Nina. “It can also allow creative expression without physical risk to an animal, and in some limited cases, may be useful for illustrative or educational purposes when clearly labelled.”

The problem with AI pet content

If AI content releases pets from the pressure to perform for clicks, then it can only be a good thing surely? Not necessarily. In fact, it could be problematic when the videos show dogs and cats enjoying situations and achieving things that would be unrealistic and even dangerous if attempted in real life.

“AI-generated content can sometimes create more pressure rather than less,” says Nina.

“When people don’t realise that content is AI-generated, they may look at it and assume that what they’re seeing is something their dog could or should do, or that the situation isn’t problematic.”

“If someone creates a video of an AI dog eating something like a corn on the cob,” she says, “there’s a real chance that viewers might assume that food is safe for dogs, when corn on the cob can cause choking and be fatal.” 

And it’s not just the physical safety of pets that may be put at risk. “While there is an element with dogs being dressed up, put on camera, or made to do things that often comes down to what an owner wants for their dog, or what they want for their social media, I’m not convinced that AI necessarily fills that gap in a healthy way,” she says.

In fact, Nina believes that there’s a wider social media pressure to consider, with people often trying to stand out or do something more eye-catching than that which they’ve already seen online. 

“There’s a risk that owners might push things further and further. What people consume online shapes their idea of what a dog is, and how a dog should live, and AI introduces a dangerous element by making dogs appear able to do anything, regardless of whether it would be good or bad for a real animal.”

Nina says that if AI content is replacing coercive or stressful content, then yes, it could provide short-term relief for the real animals that might otherwise have been put into those situations. But she worries that AI lacks nuance, including an understanding of how dogs communicate through their bodies, faces and subtle behaviours.

“Those are things AI can’t accurately replicate and losing them risks further misunderstanding rather than improving welfare,” she says. 

Long term consequences of AI pet content

Nina isn’t alone in her concern, an RSPCA spokesperson said, “AI can help us enjoy wonderful animal content without needing to use real animals for entertainment, but whilst we might enjoy the creativity of AI, it’s important we continue to protect the welfare of animals.”

“We also need to make sure that fun videos don't accidentally encourage anyone to put real animals in uncomfortable or unsafe situations by trying to copy what they see online.”

The spokesperson agrees that AI can assign animals human feelings or thoughts that aren’t accurate; known as anthropomorphism, where we attribute human characteristics or behaviour to an animal. The RSPCA spokesperson uses the example of a pet seemingly giving side-eye to explain this further.

“In a video it might be used for a joke, but in real life, a side-eye can often be an ‘I‘m trying to chill out’ signal because they feel a bit uncomfortable. And if AI shows animals behaving inaccurately, it can confuse us about how to interact with our own pets in the real world,” they said. 

The anthropomorphism that could be encouraged by AI ‘humanising’ animals is dangerous because it can lead to a misunderstanding of pets, and to inappropriate behaviour around them (such as allowing a child to get into the bed of a sleeping dog). 

Nina thinks that AI content reinforces mistaken beliefs about what animals enjoy. “It creates unrealistic expectations about what animals should tolerate and how they should behave. When those expectations are applied to real pets, it can lead to situations that are not only unfair, but potentially dangerous,” she says.  

“When AI animals can be made to do anything without consequence it risks reducing animals to playthings or props, removing a natural ethical barrier.”

The solution?

While AI-generated pet has its pitfalls, it isn’t going away, so what can we do to keep our pets safe? Nina believes AI should be used both thoughtfully and transparently and would like content that uses AI to be clearly marked. 

She says she sees many videos that are inappropriate and even dangerous, where people genuinely believe the dog shown is real and a lot of conversations where people are trying to guess whether a video is real or AI-generated. 

“The more that happens without clear labelling, the more the lines blur. Without a simple, explicit distinction between ‘this is a real dog’ and this is not a real dog,” people are left filling in the gaps themselves,” she says, “this feeds directly into unrealistic expectations, poor judgement, and a growing disconnect from what dogs’ experience and communicate in real life.”

“Clear labelling isn’t about limiting creativity; it’s about preventing misunderstanding and protecting real animals from the knock-on effects of what people think they’re seeing.”

Why we need labelling

Samantha agrees that we might need to change our relaxed approach to content. “I think as these AI tools get smarter, it will be harder to tell if it is AI or not, and there is a concern that if a trend starts of a dog doing something dangerous that people could attempt to recreate that,” she says. 

“At the moment social media platforms allow people to disclose if they have used AI to create their post, but it isn’t enforced, so this is perhaps something that will have to change.”

Nina also says it’s important to consider the cumulative impact of this content. One video on its own may seem harmless, but repeated exposure could be problematic.

“Over time, that shift in expectations affects how people relate to their pets and how much space they allow for individuality, limits, and emotional needs. Keeping real animals, with real variation and real boundaries, at the centre of these conversations is essential,” she says.