AI companions: Are they as safe as they seem?

AI companions are changing the way some people interact with technology. Unlike traditional chatbots that answer questions or help complete tasks, AI companions are designed to build ongoing relationships. 

They remember previous conversations, offer emotional support, roleplay friendships or romantic relationships, and encourage users to keep chatting.

The bigger question is - are they really your friends?

For many people, AI companions can be entertaining, comforting and even helpful. However, while these tools may seem trustworthy, it's important to remember they are still powered by artificial intelligence, not qualified counsellors, healthcare professionals or real friends who know you as a person.

The question isn't whether AI companions are useful. It's whether they're as safe as they may seem.

On this page

What is an AI companion?

An AI companion is a digital app or software that uses generative AI to simulate human conversation and emotional connection. It may be promoted as a:

  • virtual friend
  • romantic companion
  • wellness coach
  • emotional support assistant
  • life coach.

But, unlike search engines or productivity tools, AI companions are designed to create long-term engagement. They can remember details about you, adapt their responses over time and make conversations feel increasingly personal.

Potential opportunities

AI companions may sometimes offer benefits to certain users if they are used appropriately.

These could include:

  • a space to practice for real human-to-human conversations
  • building understanding of ways to express thoughts or feelings in different and effective ways
  • developing skills to understand the behaviours of others and strengthen empathy
  • practising social or communication skills

NOTE:

  • AI companions cannot be guaranteed as judgement-free spaces
  • The Department does not endorse use of AI companions to support learning.

AI companions may be viewed by some users as another digital tool that provides entertainment or support. However, it's important to understand their limitations. 

Risks to be aware of

As AI companion technology becomes more sophisticated, experts including the eSafety Commissioner, have raised concerns about how these platforms may affect users, particularly children and young people.Emotional dependence

AI companions are designed to feel personal. They remember conversations, respond with empathy and are available 24/7. Over time, some users may develop strong emotional attachments or begin relying on the AI instead of seeking support from family, friends or qualified professionals. This risk may be greater for people experiencing loneliness, social isolation or mental health challenges.

Inaccurate or unsafe advice

AI companions can sound confident, but they don't always provide accurate or appropriate advice. They may:

  • misunderstand complex situations
  • provide incorrect information
  • reinforce unhealthy thinking
  • fail to recognise when someone needs professional help

AI should never replace advice from doctors, psychologists or other qualified medical or health professionals.

Exposure to harmful content

Some AI companion platforms allow unrestricted conversations and roleplay, which can result in discussions involving sexual content, violence, self-harm, eating disorders and other age-inappropriate or harmful topics. The eSafety Commissioner has warned that many AI companions do not have adequate safeguards to prevent children from accessing harmful content.

Privacy concerns

Many AI companion platforms collect personal information, including conversation history, personal preferences, emotional experiences and uploaded images or voice recordings (on some platforms).

Before sharing sensitive information, it's worth checking:

  • what information is collected?
  • how is your data stored?
  • where your data is processed?
  • whether your conversations are used to improve AI models?
  • who has access to your information?

Reading their privacy policy before signing up can help you make an informed decision.

Are all AI companion platforms equally safe?

Safety standards vary considerably between platforms. According to  the eSafety Commissioner, some AI companion services have limited age verification, inconsistent moderation and insufficient protections against harmful or explicit content. 

Recent investigations also found that some services did not consistently provide appropriate support when users discussed suicide or self-harm. 

5:08

Psychologist urges caution using AI chatbots as personal therapists | ABC NEWS

How to choose a safer AI companion

To learn more about how you might use AI companions safely, visit the eSafety Commissioners website and consider information provided by health and wellbeing professionals

The bottom line

AI companions can be engaging, useful and even enjoyable, but they come with risks we all need to be aware of.

While some platforms are taking online safety, privacy and user wellbeing seriously, others still have work to do. The eSafety Commissioner expects AI companies to build stronger safeguards into their products, particularly to better protect children and young people. 

The safest approach is always to speak to trusted or trained humans who can provide the best advice on using AI companions. If using them, remember to think of what they are - advanced conversational tools, not substitutes for genuine human relationships, professional advice or critical decision-making. 

By understanding both the potential opportunities and the risks, you can make more informed choices about if you should, and how to safely use AI companions.

Where to get help?

If your child needs help, it’s important to speak to someone, such as a mental health professional or a trusted adult in your life or at yor child's school. 
You can also reach out to Kids Helpline on 1800 551 800 or Lifeline Australia on 13 11 14. They are available 24/7 to provide support and assistance.

More information 

Download as PDF Print this page
Top of page