Russian IT specialist Igor Bederov stresses that users should avoid sharing personal, financial, or confidential information with open AI systems, highlighting increasing risks of data breaches and unregulated use in the evolving AI landscape.
Public use of AI chatbots should not be treated as a safe place for sensitive material, according to Russian IT specialist Igor Bederov, who warned Gazeta.Ru that people should avoid entering passwords, payment details, documents and other private records into open systems. He said encryption protects information only while it is being sent to a server, after which operators, staff or contractors may be able to access it.
Bederov urged users not to share passcodes, API keys, crypto wallet seed phrases, bank card and account data, income records, medical files or commercial information with public AI services. His warning reflects a wider concern in the technology sector that employees often blur the line between personal and work use of such tools, creating what analysts describe as shadow AI: unsanctioned use of AI systems without company oversight.
That problem is not confined to individual users. TechRadar has reported that businesses are struggling to see where AI tools are being used, while staff training and clear governance policies often lag behind adoption. The result can be data leakage, compliance failures and poor decisions made on the basis of unchecked model output. The same risks become more acute as agentic AI and coding assistants gain broader access to files, systems and credentials.
The privacy debate has also been sharpened by reporting from 404 Media and Tom’s Hardware that OpenAI has used large numbers of contractors to review some ChatGPT conversations as part of model improvement work. Axios reported this week that OpenAI has also disclosed several recent AI safety incidents, underlining how difficult it remains to keep advanced systems within intended limits. Taken together, the warnings point to the same conclusion: users and companies should assume that anything they enter into a public AI system may carry more exposure than they expect.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
