Because outside of the tech community (in fact, many even inside of it), almost 100% of the folks consider what these chatgpt like tools answer as the truth without questioning it, or cross-verifying it even once.

In that case most of the mitigations listed by the author don't help though (e.g. surfacing the source). That's also no different than traditional works with citations (be it Youtube videos or peer-reviewed academic papers), where anybody rarely verifies what's written in the cited sources.

The only real alternatives would be:

- Kicking off a deep research-like investigation for each simple query

- Introducing a trusted middleman for sources, significantly cutting down the available information (e.g. restricting Wikipedia to locked-down/moderated pages)

- Not having any information at all, as at some point you can rarely every verify anything depending on how hard your definition of "verify" is