AI will still need checking for a while …

AI output usually cannot be left “as is” …

There needs to be some form of checking before letting it loose, unless incorrect information has virtually no impact. “Trust but verify” applies.

The degree of checking will vary with the risk/impact of errors.

For example, LinkedIn is apparently using AI to generate article ideas and then send invitations for people to contribute to the articles. Sometimes I get multiple invitations a day.

The results range from the rational to the irrational.

Here’s one of the latter that is more subtle:

Software Project Management • 326,911 followers
How can you improve communication with team members using problem-solving skills?

If the team members are using problem-solving skills, couldn’t they solve the communication problem themselves? 😉

Of course, it’s also unclear if the team members are using problem-solving skills, or the invitation meant using problem-solving skills to address communication … and also why checking is needed!