ChatGPT
A guide for families with key information about ‘ChatGPT’, including the age rating, key terminology, risks and instructions for enabling parental controls and safety settings.
- Part of
How ChatGPT works
ChatGPT is an artificial intelligence chatbot designed by OpenAI that:
- answers questions
- generates content
- can interact with users in a way that closely resembles a human
ChatGPT does not actually have real intelligence. Instead, using the vast amounts of text from books and websites it has been trained with, it spots patterns and generates text that fits this pattern. Because of this, it can respond to user requests in a useful and informative way. However it can make errors and share misinformation.
ChatGPT can be used by any child or young person with access to a web browser. Users do not need an account to access Chat GPT.
There are paid for versions of ChatGPT called ChatGPT Plus and ChatGPT Pro which give extended access to the most advanced models and the opportunity to test new features.
Minimum age requirements
To create an account users must be aged 13 or above. Those under 18 must have parental permission. However, there is no age verification when creating an account.
Features
Archived Chat
Chats which have been hidden from the sidebar but have not been deleted. This allows ChatGPT to retain its memories of the chat and refer to it in future responses.
DALL-E
An AI model that generates images based on text prompts. This model is incorporated into ChatGPT and can be used for free.
GPT Store
Where users can browse various Generative Pre-trained Transformers (GPTs), which are custom applications built using AI. These can range from writing assistants to personality tests.
History
Records of previous chats retained by ChatGPT. The chatbot can refer to information from these in its responses.
Memory
ChatGPT can remember details between chats, allowing it to give personalised and relevant responses.
Plugin
Extra tools that can give ChatGPT additional features. For example, they can enable ChatGPT to book tickets, shop online and count calories.
Search
ChatGPT can search the internet, allowing it to incorporate real-time information into its responses instead of relying on the information it learned during training. It includes a link to the source, allowing users to see where the information comes from.
Temporary Chat
A chat that is not saved permanently and is deleted after the session ends. Data is not collected from temporary chats.
Voice Assistant
Users can use their voice to input a prompt, rather than type it. This is only available on paid-for plans.
Terms
Chatbot
Short for ‘chat robot’. A type of AI that responds to questions using a chat function.
Hallucination
When AI unintentionally makes up information that isn’t true but presents it as fact.
Prompt
Prompts are the questions that users ask ChatGPT.
What ChatGPT offers
Learning and education
ChatGPT can help young people with their education and schoolwork. It can:
- explain difficult concepts in a simple way
- quiz them to test their knowledge
- help them plan study schedules
While ChatGPT can support a young person’s learning, they must not use it to do their work for them.
Creativity
Young people can use ChatGPT to nurture their creativity and find inspiration. For example, they can ask for story ideas or artistic prompts. Additionally, it can be used to play games that encourage imagination, such as roleplaying games in which ChatGPT plays the part of gamemaster.
Productivity
When used correctly, AI can boost productivity. It can make researching, problem-solving and generating ideas easier.
ChatGPT can also speed up everyday tasks such as creating schedules or budgets, or writing messages and emails.
Managing risks
Misinformation
ChatGPT will sometimes confidently present made-up information as a fact. If your child is exposed to this misinformation, they could become confused and use this false information, potentially harming their schoolwork and knowledge on topics.
These actions can reduce the impact of misinformation.
Verify information
Encourage your child to verify any information they receive from ChatGPT. They can do this by checking the chatbot’s response using trustworthy sources like educational websites, credible news sites or fact-checking websites such as Full Fact.
Develop critical thinking skills
Helping your child develop their critical thinking and media literacy skills will improve their ability to recognise misinformation. Teach them to stop and question any information that they see and analyse this information rather than taking it at face value.
Content and use
Inappropriate content
ChatGPT has guardrails in place, which are in-built safeguarding and content moderation tools intended to prevent the chatbot from responding with harmful, biased, dangerous or inappropriate information. However, these guardrails are not perfect, and inappropriate content can slip through into the chatbot’s responses.
Inappropriate use
If your child uses the chatbot in the wrong way, it can harm them. For example, if your child uses ChatGPT to do their schoolwork for them, they will not learn anything and will likely be caught plagiarising.
ChatGPT can impact a child’s wellbeing if they go to the chatbot looking for advice on their mental health, something the AI is not equipped to handle. This could also lead to oversharing, as when the child looks to confide in ChatGPT, they may share personal information that is then collected by OpenAI.
Steps to help your child use ChatGPT safely and reduce the risk of it generating inappropriate content
Set up parental controls
Using ChatGPT’s parental controls, you can reduce sensitive content, giving your child additional protections from graphic content. It will also alert you to any safety concerns in your child’s conversations, allowing you to intervene if they are inquiring about inappropriate topics.
Only ask safe questions
ChatGPT is a user-driven app, with the chatbot only sharing information and content in response to the user’s input. Because of this, the user has a lot of influence on the type of content included in the responses they receive. Encourage your child to only ask safe, educational questions and refrain from asking questions about violent or sexual topics.
Supervise AI use
When your child first starts using ChatGPT, try using it alongside them. Doing so will give you understand the type of questions they ask the AI and allow you to see if they have the critical thinking skills required to use the AI safely.
Use the reporting tool
Show your child how to use the report feature on ChatGPT. This can be accessed in-browser by selecting the 3 dots in the top corner of the screen. Using this tool will make it less likely that ChatGPT will share inappropriate content in future, making the service safer for everyone.
Discuss oversharing
Discuss what your child can and can not share online, to help them avoid giving away sensitive personal information. Explain that they should not tell ChatGPT their address, school name or any other information that could be used to identify them and increase the risk of their identity or data being stolen.
Agree on boundaries
Agree on boundaries with your child about when and how they use ChatGPT. Explain that they shouldn’t use ChatGPT as a therapist or to complete work for them. Instead, show them how they can use the chatbot as a tool for learning and creativity. This can also help prevent overreliance on the chatbot.
Settings to protect your child
If your child wants to use ChatGPT, the following settings can make their experience with the chatbot safer.
Set up parental controls
OpenAI has introduced a range of parental controls for ChatGPT which allow you to link your account to your child’s and then monitor and manage their use of ChatGPT. Using parental controls, you can reduce sensitive content, disable image generation and set quiet hours when ChatGPT cannot be used.
Setting up parental controls
- Create or login to your ChatGPT account.
- On the homepage, select your profile picture.
- From the menu, select Settings.
- Inside Settings, select Parental controls.
- Select Add family member.
- Type in your child’s email address, then select My child.
- Select Send.
- Once your child accepts your request through their email, you can manage their account.
Adjust privacy settings
OpenAI automatically collects user data and chat history to train the artificial intelligence and improve the performance of ChatGPT. To improve data security and privacy, users can opt-out of having their data collected and delete their chats.
Opting out of data training (on the web)
- Create or login to your ChatGPT account.
- On the homepage, select your profile picture.
- From the menu, select Settings.
- Inside Settings, select Data controls.
- Select Improve the model for everyone.
- Turn off the toggle next to Improve the model for everyone.
Opting out of data training (on the app)
- Select the 3 horizontal lines in the top left of the screen.
- Tap your profile name at the bottom of the menu.
- Select Data Controls.
- Turn off the toggle next to Improve the model for everyone.
Clearing chat history (on the web)
- Select your profile picture.
- Select Settings.
- Select Data controls.
- Go to Delete all chats and select Delete all.
- Select Confirm deletion.
Clearing chat history (on the app)
- Select the 3 horizontal lines in the top left of the screen.
- Select your profile name.
- Tap Data controls.
- Select Clear Chat History.
- Select Clear Chat History again.
Review reporting feature
If ChatGPT responds to a prompt with inappropriate or misleading content, it is important that the user reports the chat. This will make it less likely that ChatGPT will share this kind of content in future and make the chatbot safer for your child and other users.
- Inside the chat, select the three dots in the top right of the screen.
- Select Report, next to the flag icon.
- Select the reason for your report and select Next.
- Choose the option that best fits your reason for reporting and select Next.
- If you want, provide more details and select Submit.