I believe, if this sort of generative AI is going to be trustworthy in the future, we need some sort of external verification system so we can make our own trust judgements based on the data used to train the system. For example, if a system is trained including 4chan as a data source, I'm going to trust it less than if it wasn't trained using that source.
I don't think big business yet realises how important the training data is but, as soon as they do, they will want the AI companies to provide guarantees about the sanity and appropriateness of the training data.
Is it a relatively new expectation of society that content should be indefinitely preserved? Prior to the widespread adoption of VHS I wouldn't have thought it would be in the public consciousness.
I'd class myself on the 'keep all the things' side of any preservation argument as building on our cultural history can only be a good thing (artistically speaking - I don't care much for the capitalist tears regarding copyright infringement). Yet I find it interesting, amusing even, that we expect, by default, that content should be preserved and readily available. In 100 years we'll still be able to view any episode or 'Mrs Brown's Boys' yet have no way to see Shakespeare's premiere of 'A Midsummer Night's Dream'.
For clarity, in 100 years I still will not have watched even 1 episode of 'Mrs Brown's Boys'.
Malaysia halts music festival after same-sex kiss by UK band The 1975 (www.reuters.com)
ChatGPT Can Be Broken by Entering These Strange Words, And Nobody Is Sure Why (www.vice.com)
Reddit usernames like ‘SolidGoldMagikarp’ are somehow causing the chatbot to give bizarre responses.
deleted_by_author
Michael Gove claims work for no-deal Brexit made UK ‘match fit’ for pandemic (www.theguardian.com)
Remember what "subscription" used to mean?
The article about the “subscription” HP ink made me realise something....
Describe a game in 5 words or less, and see if anyone can guess it.
I'll go first. Deranged Spiderman with guns