Artificial intelligence - AI chat - Stay Safe Partnership

provide parents and carers with information about the function of AI chatbots
outline the benefits and potential risks they may present to young people
offer advice on ways to keep young people safe

Artificial intelligence (AI) chatbots are conversation engines. They are trained to provide human-like responses and dialogue using a process known as natural language processing (NLP). They can interpret human language and respond on their own to messages they have been sent. Chatbots learn from all the data that is available to them. They are only as good as the data they are trained on. AI chatbots can be found in various applications, including:

messaging apps (for example, My AI on Snapchat)
educational platforms
customer service websites

They can provide information, answer questions, and even engage in playful banter.

Examples of AI chatbots:

chat GPT (ages 18 and over, or 13 with parental consent) - generates human-like responses to text-based conversations
character AI (13 and over) - a chatbot that generates human-like responses based on the user’s customisation
HiAi (17 and over) - allows users to create a character to talk to
Replika (18 and over) - an AI companion that is paired with a 3D avatar
Ai Friends (17 and over) - has pre-set fictional characters for the user to chat with

AI chatbots are becoming increasingly popular tools for communication, learning, and entertainment. While they can offer many benefits, it’s essential for parents to understand the potential risks, especially when it comes to children’s interactions with such technology. Some have been listed below.

Benefits:

language skills - conversing with chatbots can help children practice language and communication skills
educational support - chatbots can assist with homework, answer questions, and provide resources for learning
24 hour availability - chatbots can offer support anytime, making them accessible for children who may need help outside of regular school hours
engagement - many chatbots are designed to be fun and interactive, making learning more enjoyable
entertainment and creativity - AI chatbots can provide prompts for games like charades or drawing challenges. They can provide inspiration for writing stories and creating quizzes, riddles and puzzles

Potential risks of AI chatbot use to children:

misinformation and inappropriate content - AI systems are not perfect. They can misinform, be inappropriate in the conversation they make, and their information can be inaccurate and biased
sexualised content - some chatbots have received reports of sexual harassment and complaints for initiating sexualised dialogue with users. Whilst most AI chat apps are free to download, they can include popup adverts. We found that these adverts were often for live chat apps or websites that were not age appropriate for the rating they were given
negative influence - children can be poorly advised by this technology, which can lead to serious and dangerous consequences. Chatbots can disclose harmful information which can influence vulnerable users to put themselves at risk
privacy concerns - children may unknowingly share personal information, leading to potential privacy violations. Large language models can learn from data shared with chatbots and may use information shared in future to respond to other people’s queries. Not all AI chatbots are completely secure so data shared could be leaked. Many AI apps collect your personal data, including children's online activity and interactions, which can be used for target advertising, profiling, or even sold to third parties
social isolation and damage to mental health - excessive interaction with chatbots might lead to reduced face-to-face social skills and interactions. Talking with a chatbot can become problematic when a young person stops speaking to trusted adults or accessing services such as counselling when facing mental health challenges
emotional vulnerability - children may develop attachments to chatbots, which could impact their emotional health and understanding of real-world relationships

Tips for parents and carers:

know your tools - research the chatbot and test it out before allowing your child to use it, understanding its features and potential risks
set boundaries - establish screentime limits and guidelines for when and how your child can interact with chatbots. Agree that your child can only use chatbots when a trusted adult is present, discuss their experiences and emphasise the importance of human interactions
discuss privacy - emphasise the importance of not sharing personal information online. Teach children what information is safe to share
fact-check together - teach critical thinking skills to help your child understand that not everything a chatbot says is correct. Encourage them to double-check information with reliable sources
open communication - ask your child about their experiences and feelings when using chatbots. Encourage them to ask questions and raise concerns. Try to avoid any judgement or dismissiveness that could dissuade them from sharing
talk to others - speak to your child's teachers and other parents and carers to discuss experiences and share advice about navigating technology with children