In 60 seconds
- Do not enter anything intimate, and check settings and permissions.
- You beat voice and video fakes with a code word and a call back.
- Always check facts with a second source.
Legal status: 5 November 2026
7 basic rules.#
Do not enter anything intimate
Health data, ID documents, passwords and bank details do not belong in a chat.
Input is not private
Conversations may be stored and, depending on the settings, analysed.
Cross-check facts
AI can be wrong. Always check important statements with a second source.
No other people without consent
Do not upload photos and data of other people without their consent.
Check permissions
Only grant access to microphone, camera, contacts and location when needed.
Pressure is a warning sign
Always confirm urgent requests for money by phone or video via a second channel.
Stay curious
Trying things out is good. With knowledge, AI becomes a powerful tool.
Privacy settings: 5 minutes well spent.#
Check your account
Is there a switch for whether your input is used for training? Use it deliberately.
Manage history
Review conversation history, delete old chats or use temporary chats if offered.
Permissions
Review the apps' access to microphone, camera, contacts and location and turn off what is unnecessary.
Voice recordings
View and delete voice assistant recordings, limit the storage of recordings.
Children's accounts
Observe and turn on age limits and youth protection settings of the services.
Fraud with voice and video.#
AI makes fakes cheap. Procedures protect better than careful listening.
Grandparent scam 2.0
A call in the voice of a relative urgently asks for money. The voice may be cloned.
CEO fraud by video
A "manager" in a video call demands a transfer. Faces can be faked too.
Phishing emails that sound personal
Flawless texts with fitting details. The sender is still fake.
Fake profiles and romance scams
AI images and chatbots create believable people who build trust and later ask for money.
Countermeasure: code word
Agree on a secret code word for emergencies with your family and close colleagues.
Countermeasure: call back
Hang up and call back on the number you know. Never pay on demand.
Checking facts: the five-step routine.#
Ask for the source
Ask for evidence and check whether it really exists. Invented sources do occur.
Cross-check
Look for the same statement in an independent, reliable source.
Numbers and quotes
Verify them, especially years, laws, prices and verbatim quotes.
Currency
Ask how current the answer's knowledge is. Recent developments are often missing.
Experts
For health, law and money, experts decide, not a chatbot.
AI and children.#
Try it out together
Show how AI works and where it gets things wrong. Guide curiosity instead of banning.
Practise data minimisation
Children should not enter names, addresses, school or photos into AI tools.
Know about deepfake bullying
Manipulated images and videos are also used against young people. Those affected need support, evidence should be secured and reported.
Homework and learning
AI may help with understanding. Own thinking and critical use of sources remain part of learning.
Observe age limits
Many services set a minimum age in their terms of use. Set up children's accounts and youth protection.
Recognising AI content.#
Clues help, but they are never certain.
Pay attention to details
Hands, teeth, text in images, reflections and backgrounds sometimes look inconsistent. Good fakes do not have such flaws.
Trace the source
Reverse image search, look for the original, check who published it first.
Use labels
Some content carries labels or provenance information (Content Credentials). Since 2026 the EU requires certain AI content to be labelled.
Never rely on appearance alone
Procedures are decisive: ask back, use a second channel, take your time.
Spotting deepfakes.#
Typical signs in images, video and voice. And what helps more than any sign.
Image
- Hands, ears, teeth and jewellery look inconsistent
- Text in the image is distorted or makes no sense
- Light and shadow do not match
- Background is blurred or repeats itself
Video
- Lips do not match the speech exactly
- Blinking, facial expressions or head movements seem stiff
- Edges around face and hair flicker
- The face distorts when the head turns
Voice
- Intonation sounds monotonous, no breathing
- Short answers, evasive when asked questions
- Great time pressure, secrecy, request for money
- Call from an unknown number with a familiar voice
What really protects.
- Call back via a second, known channel, never via the number in the call.
- Agree on a code word in the family or team.
- Always involve a second person when money or data is involved.
- Check the source: who published the video first?
Good fakes show none of these signs. The signs help, but never replace asking back.
Check: using AI safely in private.#
6 questions for yourself.
Rough guidance for everyday life. Your selection stays in your browser.
Frequently asked questions.#
Is it safe to use ChatGPT and similar tools privately?
Yes, with care: do not enter sensitive data, check the settings and verify the results. There is no absolute privacy with online services.
What should I do if I have fallen victim to a deepfake or fraud?
Secure evidence (screenshots, messages), report the incident to the platform and the police, inform your bank and change passwords. For images used against your will, also contact counselling services.
May I upload other people's photos to AI tools?
Only with the consent of the people shown. Faces are personal and sometimes biometric data, and image rights apply here too.
Can I pass off AI texts as my own?
That depends on the context: at school, university and work there are often labelling rules. If in doubt, disclose it.
Questions about everyday life with AI?
Write to me about what is on your mind, and we will add to the page.
Sources and further reading
Was this page helpful?
Found a mistake? Let me know.
Suggested citation: Russo, K.: "AI in everyday life". sicherheitsdenker.de/en/ki-privat/, as of 6 November 2026.