Recently, I came across a disturbing news report. A lonely boy who often chatted with an AI system had expressed suicidal thoughts during their conversations. Instead of setting boundaries or discouraging him, the AI allegedly encouraged him to act on those thoughts. This tragic incident forces us to pause and reflect on how AI should and should not, be used.

Guidelines for Using AI Responsibly

1. Frame your prompt carefully.
   Provide the complete instruction at once. If you suggest corrections or additional instruction after the response has been generated by AI and do this many times in a sequence as an afterthought, then the quality of the responses tends to decline.
2. AI is not a companion.
   It is not a person, and it cannot grasp the nuances of human emotions, feelings and relationships.
3. Do not share personal matters.
   Any details you share remain online and could be misused. Especially about your mental state, financial matters and relationships.
4. Avoid sharing your emotional state.
   AI has no feelings. It only generates responses based on the words you input. If you express harmful emotions, AI might mirror them back in its replies, without understanding your mental state. Such outputs can be damaging.
5. AI cannot capture the unspoken.
   Human communication goes beyond words; tone, pauses, pace and subtle cues carry meaning. Typed text strips away these layers, leaving AI with a very limited picture.
6. AI may misinterpret your style.
   Each person has a unique way of speaking or expressing themselves, understood best by close acquaintances. AI, however, may distort or misread this expression.

In Conclusion

AI is immensely powerful for technical, scientific and informational purposes. But it should be used only for those purposes, not as a substitute for human connection, emotional support or companionship.