Home > Resources > Toolkits > Artificial Intelligence: Risks & Safety Tips for Survivors

Artificial Intelligence: Risks & Safety Tips for Survivors

Many reasons might lead someone experiencing abuse to choose an AI tool. It may feel easier than explaining your situation to another person. It can be free or low-cost, available 24/7, and accessible from home. For someone living in a small or rural community, it may also feel like one of the few options available. 

You are not doing anything wrong by using AI. AI can sometimes help you find general information, organize your thoughts, draft questions, or understand unfamiliar terms. 

At the same time, using AI tools carries risks. They may collect or retain information you enter, provide incorrect information, give advice that does not apply where you live, or respond in ways that are not appropriate for your situation. 

The goal of this information sheet is to explain these risks so you can decide whether to use AI, and to share some tips for using AI safely if you do decide to use it.  

Potential Risks of Using AI

1. Your AI conversation may not be as private as you think 

An AI chat is not like a private conversation with a lawyer, doctor, counsellor, or support worker. These professionals are legally required to keep the information you share with them confidential in most circumstances.  When you use an AI chatbot, the information you enter is being sent to a company or service provider. Depending on the tool and its settings, conversations may be stored, reviewed, retained for a period of time, sold, disclosed in a subpoena request, or used in other ways described in the company's privacy policies. This means other people may be able to access the information you enter without you knowing. 

2. The information you provide might be stored longer than you think or permanently 

Deleting a conversation from your screen does not necessarily mean that every copy of the information has immediately been deleted from the company's systems. Different AI services have different retention and deletion practices. Some may retain information for a certain period of time for security, legal, technical, or other purposes. There may also be circumstances where a company is legally required to preserve or disclose information. For example, information held by a company could be subject to a legal demand such as a court order or subpoena, depending on the circumstances and jurisdiction.

3. AI can provide inaccurate information

AI can sound confident, even when it’s wrong. AI can: 

  • Give outdated information 
  • Confuse different provinces, territories, or countries 
  • Mix different areas of law or services provided by different organizations 
  • Misunderstand the facts of your situation 
  • Provide information that sounds legally correct but is not 
  • Make up cases, legislation, legal websites, or citations 

Decisions or actions based on inaccurate information can be particularly risky when you are dealing with safety, housing, family law, immigration, child protection, criminal law, or financial issues. 

4. AI cannot replace human expertise and support

It can feel easier to talk to AI. If you have experienced abuse, you may have good reasons for wanting support that feels private, immediate, and available on your schedule. An AI chatbot may feel easier because you can write what happened without worrying about a real person’s immediate reaction. It can also respond immediately when a counsellor isn't available. Those benefits are real. 

But AI cannot replace the role of a trusted human. Human support offers things AI cannot reliably provide. A counsellor, advocate, lawyer, doctor, or trusted person can: 

  • Listen to you and understand the context of your life 
  • Ask questions when something isn't clear 
  • Notice changes in your emotions or circumstances 
  • Notice physical changes related to your emotions 
  • Help assess immediate safety concerns 
  • Recognize when a situation requires urgent or specialized support 
  • Connect you with local services 
  • Understand the laws and resources where you live 
  • Help you make decisions without taking control away from you 
  • Provide accountability and ongoing support 
  • Build a relationship based on your needs and circumstances 

AI does not have lived experience, professional responsibility, or the ability to understand and navigate human relationships. 

5. Be aware of developing a relationship with an AI chatbot

AI can feel surprisingly personal. Some AI chatbots are designed to be conversational, supportive, and responsive. They may remember information, use a warm tone, ask questions, and respond as though they understand how you are feeling. 

If you are isolated or recovering from an abusive relationship, it may feel comforting to have something supportive available to talk to at any time. You may begin to feel attached to the chatbot or think of it as a friend, counsellor, or someone who understands you. 

These feelings are understandable. But it is important to remember that an AI chatbot is not a person and does not have a relationship with you the way another person does. 

Relying heavily on an AI chatbot can create risks, including:  

  • Increasing your isolation from other people 
  • Reinforcing unhealthy relationship patterns  
  • Agreeing with you automatically rather than providing other perspectives or ideas  
  • Encouraging you to rely on it more than you intended  
  • Giving advice without understanding the consequences or full picture of your safety needs and risks 

 AI can be one tool for finding general information or helping you think through questions, but it should not become your only source of emotional support or safety advice. Where possible, stay connected with people who can support you.  Visit www.ShelterSafe.ca for an anti-violence program in your community.  

Safer AI Habits 

Here are some safety tips that can help you use AI more safely and protect against the risks listed earlier in this document:

1. Share less 

Ask yourself: "Does the AI need to know this information to answer my question?" If not, leave it out.

2. Don’t include identifying details 

Avoid entering information that could identify you or someone else, such as:  

  • Your full name 
  • Your address or location 
  • Your phone number or email address 
  • Your children's names, schools, or locations 
  • Your partner or former partner's name 
  • Your workplace 
  • Your case or file number 
  • Your financial information
  • Passwords or account information 
  • Government identification numbers 
  • Medical or mental health information 
  • Detailed descriptions of your abuse that could identify you 
  • Copies of legal documents containing identifying information 
  • Photos, screenshots, or documents that contain personal information 

Instead, use: 

  • "My former partner" instead of their name 
  • "My city" instead of your address 
  • "My child" instead of your child's name 
  • "My workplace" instead of your employer's name 

For example:  

Instead of: "My name is Sarah. I live at 123 Main Street and my former partner John Smith has been stalking me..." 

Try: "I am experiencing unwanted contact from a former partner. What general safety considerations should someone in this situation think about?"

3. Double-check the details of the AI’s response 

When reading a response from an AI chatbot or search tool, ask yourself: “Where did this information come from?” 

Look for the source, such as an official government website, legislation, court website, or reputable legal organization. 

Then check: 

  • Is this information for Canada? 
  • Is it for my province or territory? 
  • Is it current? 
  • Does it apply to my situation? 
  • Can I find the same information in the original source? 

If the AI gives you a link or citation, don't assume it is real or accurate. Open the source and check it yourself. 

This is especially important when you are looking for information on legal rights, safety, housing, immigration, finances, health, or children.

4. Maintain human connection & set boundaries on your AI use 

Other humans can provide support, empathy, and assistance in ways that AI cannot. AI can be a useful outlet, but it shouldn’t replace conversation or connection with other humans.  

Here are some tips for setting boundaries around your relationship with AI: 

  • Remember that the chatbot may not always remember information accurately. 
  • Don't assume that the chatbot understands your feelings in the same way a person would. 
  • Don't rely on the chatbot as your only source of emotional support. If you are not comfortable speaking with someone you know, calling a confidential crisis line can be a good alternative. 
  • Keep connections with trusted people whenever possible. 
  • If the chatbot encourages you to withdraw from people in your life, treat that as a warning sign. 
  • If a chatbot tells you that you only need it and don't need other people, step back. 
  • Pay attention to when and how you use it. Take steps to reduce your use or reach out to other people if you find yourself spending more and more time talking to it. 
  • Take breaks from the AI conversation and reconnect with people, activities, and supports in your life. 

A helpful question to ask yourself: "Is this tool helping me connect with my life, or is it becoming a replacement for my life?" AI can be one tool in your support system. It should not become your entire support system. 

5. Seek professional support as well

Sometimes professional support isn’t immediately available. If AI is one of the tools you use while you are waiting for professional support, that does not mean you are doing something wrong. Keep in mind that AI can provide general information and conversation, but it cannot replace professional support and advice.  

If you find yourself turning to AI for support regularly, take that as a sign to reach out to professional support as well. You could call a confidential crisis line for immediate support, reach out to a shelter near you via ShelterSafe.ca, or contact a support worker or counsellor you have connected with before. 

Consider shifting the way you use AI. Rather than turning to it for answers, use it to help you prepare for your conversations with professional supports. For example, you could ask a chatbot: “What questions might someone experiencing domestic violence want to ask a family lawyer about separation and safety?” This can help you feel more confident and ready to reach out to professionals, rather than relying solely on AI. 

6. Trust your instincts & don’t let AI make important decisions for you

If an AI response makes you uncomfortable, confused, pressured, or less safe, you can stop using it. You don't have to follow its advice. 

AI can help you think through options. You remain the person who decides what is right for you. 

Technology-Facilitated Gender-Based Violence (TFGBV) is part of a continuum of violence that can be both online and in-person. If you or someone you know is experiencing TFGBV, you are not alone. You can use ShelterSafe.ca to find a shelter/transition house near you to discuss options and create a safety plan. You don't need to stay in a shelter to access free, confidential services and support.