ChatGPT was the very first AI chatbot to gain global recognition. Since its launch, many other capable AI chatbots have emerged, providing a wider range of options to meet your specific needs. AI chatbots have become extremely popular and useful tools for obtaining information, advice, and assistance on various topics. You can use them to create a business plan, plan your garden, write articles or code, compose emails, generate art, images, videos, and pretty much anything you can imagine.
However, the more advanced and integrated these AI assistants become in our lives, the more cautious we must be about sharing personal information with them. Why? Because they cannot be trusted with sensitive data.
Behind the friendly chatbot
In order to understand the privacy risks associated with AI chatbots, it's important to know how they work. These chatbots gather and store the full transcripts of conversations that take place when you interact with them. This includes all the questions, prompts, messages you send to the chatbot, and the chatbot's responses. The companies behind these AI assistants analyze and process this conversational data to train and improve their large language models.
Think of the chatbot as a student who takes notes during class. The chatbot writes down everything you say verbatim, so the full context and details are captured. The AI company then reviews these "notes" to help the chatbot learn, much like a student would study their class notes to increase their knowledge. While the intent is to enhance the AI's language understanding and dialogue abilities, it means your raw conversational data- which could potentially include personal information, opinions, and sensitive details that you disclose - is being collected, stored, and studied by the AI companies, at least temporarily.
If you're on the Internet, your personal data is already all over the place. Do you want to see where it is? Use Bitdefender Digital Identity Protection to locate and manage your personal information online.
The risks of oversharing
When you share personal or sensitive information with an AI chatbot, you lose control over where that data goes or how it may be used. AI chatbots store data on servers, which can become vulnerable to hacking attempts or breaches. These servers hold a wealth of information that cybercriminals can exploit in various ways. They can infiltrate the servers, steal the data, and sell it on dark web marketplaces. Additionally, hackers can use this data to crack passwords and gain unauthorized access to your devices.
Any data you provide could potentially be exposed, hacked, or misused, leading to identity theft, financial fraud, or the public exposure of intimate information you would rather keep private. Protecting your privacy means being selective about the details you disclose to AI chatbots.
What not to tell a chatbot
To protect your privacy and personal information, be selective about what you share with AI chatbots.
Be extremely careful with these types of data:
1. Personal Identifying Information: Avoid sharing key pieces of personal identifying information such as your full name, home address, phone number, date of birth, social security number, or other government ID numbers. Any of these can be exploited to impersonate you, leading to identity theft, financial fraud, or other criminal misuse of your personal details.
2. Usernames and passwords: Never share passwords, PINs, authentication codes, or other login credentials with AI chatbots. Even providing hints about your credentials could help hackers access your accounts.
3. Your financial information: You should never share any financial account information, credit card numbers, or income details with AI chatbots.
You can ask them for general finance tips and advice, broad questions to help you budget, or even tax guidance, but keep your sensitive financial information private, as it could easily lead to your financial accounts and assets being compromised.
4. Private and intimate thoughts: While AI chatbots can serve as a sympathetic ear, you should avoid revealing deeply personal thoughts, experiences, or opinions that you wouldn't feel comfortable sharing publicly. Anything from political or religious views to relationship troubles or emotional struggles could be exposed if conversational logs are hacked or mishandled.
5. Confidential work-related information: If you work with proprietary information, trade secrets, insider knowledge, or confidential workplace data of any kind, do not discuss this with public AI chatbots. Avoid using AI chatbots to summarize meeting minutes or automate repetitive tasks, as this poses the risk of unintentionally exposing sensitive data or violating confidentiality agreements and intellectual property protections of your employer.
A Bloomberg report highlighted a case where Samsung employees used ChatGPT for coding purposes and accidentally uploaded sensitive code onto the generative AI platform. This incident resulted in the disclosure of confidential information about Samsung, prompting the company to enforce a ban on AI chatbot usage.
Major tech companies like Apple, Samsung, JPMorgan, and Google have even implemented policies to prohibit their employees from using AI chatbots for work.
6. Your original creative work: Never share your original ideas with chatbots unless you're happy to have them potentially shared with all other users.
7. Health-related Information:
A survey conducted by health tech company Tebra revealed that:
- 1 in 4 Americans are more likely to talk to an AI chatbot instead of attending therapy.
- Over 5% of Americans have turned to Chat GPT for a medical diagnosis and followed its advice.
To protect your health data means to protect your access to proper medical care, maintain confidentiality, and safeguard against potential privacy breaches or misuse of sensitive medical information. So, never disclose your medical conditions, diagnoses, treatment details, or medication regimens to AI chatbots. Instead, discuss with qualified healthcare professionals in a secure and private setting.
How to use chatbots safely
Here are 3 things you can do to safely use chatbots and protect your privacy.
1. Be cautious about the information you provide.
2. Read privacy policies and look for chatbot privacy settings.
3. Use the option to opt out of having your data used for training language models when available.
In general, using incognito/private modes, clearing conversation history, and adjusting data settings are the main ways to limit data collection by AI chatbots. Most major AI chatbot providers offer these options.
Here are some examples:
OpenAI (ChatGPT, GPT-3):
- Use incognito/private browsing mode
- Enable "Don't save history" in the settings
- Clear your conversation history regularly
Anthropic (Claude):
- Enable the "Privacy filtering" setting to prevent data collection
- Use the "Incognito Mode" feature
Google (Bard, LaMDA):
- Use guest mode or incognito mode
- Review and adjust your Google data settings
Next level: Make AI keep you safe
While being cautious is wise, what if you could take privacy protection even further by leveraging an AI assistant explicitly designed to safeguard you?
Scamio is our next-gen AI-powered chatbot, designed to detect scams and fraudulent activities. You can send Scamio the tricky text, email, instant message, link, or even QR code you received, and it will provide an instant analysis to determine whether it's a scam attempt.
To start a conversation, visit scamio.bitdefender.com/chat or chat directly with Scamio on Facebook Messenger.
