AI Writes Well But Users Must Take Responsibility for Errors

Noam Cohen highlights the growing use of artificial intelligence (AI) for drafting messages, emphasizing that while AI can produce polished text, the ultimate responsibility lies with the human user. He recounts receiving a well-written email from a client that was entirely generated by AI, yet signed by the client who did not fully stand behind its content. Cohen stresses that despite AI handling the writing, the user must carefully review and approve the final message before sending it, as any mistakes will be attributed to them, not the AI or its developers.

Cohen outlines five critical questions to ask before sending AI-generated content: whether the user has thoroughly read the text, verified factual details like dates and names, ensured the tone matches their voice, considered how the recipient will perceive it, and confirmed that the decision to send remains theirs. He warns that AI can confidently fabricate details and produce overly polished language that may feel impersonal, so users should personalize and validate content to maintain authenticity.

He concludes that the key skill in the AI era is not just using AI to write but understanding and owning the output. Cohen, founder of CritiqeIL, an AI consulting firm in Israel, advises that users must retain accountability for their communications, leveraging AI as a tool without relinquishing control or responsibility.