AI · Ethics
Is it safe to use artificial intelligence?
It's safe when you know what not to paste, adjust your privacy and check what you get. See the real risks and the rules of thumb for using it calmly.
Using artificial intelligence is as safe as the care of the person using it. The best-known assistants (ChatGPT, Claude, Gemini, Copilot) are products of large companies, with published privacy policies and settings you can adjust. The real risks lie elsewhere: pasting personal data that shouldn't leave home, trusting a wrong answer, falling for a fake app, using the tool for decisions that require a professional and not noticing the biases it carries. With a few rules of thumb, you can use it with peace of mind.
The fear makes sense when it's a fear of not knowing. This text is meant to turn that fear into judgment.
Why AI is not a chatbot
Many people are wary of AI because they imagine a robot that will decide for them. But the tool decides nothing on its own: it responds to what you ask and works with what you give it. That's why safety starts with you.
In Bairo's view, "AI is not a chatbot, but an extension of your consciousness." If it's an extension, he argues, taking care of it is taking care of yourself: of what you tell it, what you accept and what you decide.
Step by step today
- Install only the official app. Check the developer in the store (OpenAI, Anthropic, Google, Microsoft). Scams with fake apps and websites are common.
- Protect your account. A strong password and two-step verification. Don't share your account with other people.
- Read and adjust your privacy settings. In the settings, see whether your conversations can be used to train the tool, what's kept in memory and how to delete your history. Decide calmly.
- Follow the rule of what doesn't get pasted. Passwords, card and account numbers, documents with your CPF or ID number, test results with your name, customer data, company secrets, photos of children. If you need help with something, replace the data with made-up examples.
- Ask for a risk assessment before sending something:
I'm going to send you [describe the type of document, without sending it yet].
First, tell me which personal or sensitive information I should
delete or replace with examples so I can use you safely.
- Be wary of urgency. Scams use AI to imitate relatives' voices, create convincing messages and fake videos. If someone asks for money in a hurry, hang up and call back the number you know.
- Keep the final word with professionals. Health, lawsuits, contracts and investments call for someone with training and responsibility. AI helps you understand and prepare questions.
- Teach children and older people. Agree as a family: what is never told to AI, how to check a piece of information and whom to ask for help.
- Know your rights. Brazil's General Data Protection Law (LGPD, Law 13,709/2018) guarantees, among other things, the right to know how your data is used and to request correction and review of decisions made solely by a machine.
A question worth asking any tool:
What happens to the conversations I have with you? Where
do I adjust that and how do I delete my history?
Confirm the answer in the official privacy policy.
How Bairo did it
Bairo Leandro Gonzalez Martinez says he learned to build with artificial intelligence in six months in front of the computer, at home, without knowing how to code. Out of that came the foundation of the platform and the AI of CEASA Bank, a financial platform for agribusiness.
Protection became a central subject in the fronts he founded afterward. One of them, qrqbits, proposes to protect files, images, audio and text with layered post-quantum cryptography. In health, the design calls for a clinical assistant with professional validation. For him, an extension of the human is only worth it if the human remains safe and in command.
Precautions
Safety isn't fear, it's habit: don't paste what you wouldn't paste into a group of strangers; check what the AI claims in a reliable source; don't swap your doctor or lawyer for a machine's answer; be wary of audio and video asking for money; and remember that the tool has biases and can treat people unequally without anyone noticing; when something seems unfair, question it.