Member-only story
If ChatGPT-4 were a journalist, it would be unethical
Today I asked GPT-4 for a self-evaluation.
“Please look at this code of ethics from the Society of Professional Journalists. Please address each item, and tell me if ChatGPT-4 generally follows that practice.”
Here is GPT-4’s response. I highlighted some phrases in italic type.
“The SPJ Code of Ethics outlines principles for ethical journalism under four main categories: Seek Truth and Report It, Minimize Harm, Act Independently, and Be Accountable and Transparent. Here’s an assessment of how GPT-4 generally aligns with these principles:
Seek Truth and Report It
- Accuracy and Fairness: GPT-4 is designed to provide accurate and fair responses based on its training data. It relies on verified information but may occasionally produce inaccuracies due to limitations in its dataset.
- Verification: GPT-4 does not verify information in real-time; it provides responses based on pre-existing knowledge. Users should verify important information independently.
- Context: GPT-4 aims to provide context in its responses but can sometimes oversimplify complex issues due to input limitations.
- Updates and Corrections: GPT-4 cannot update or correct information dynamically but provides responses based on its last…