There is so much disinformation in the world right now.
-
There is so much disinformation in the world right now.
And in recent years, all of it gets amplified by AI web scrapers that eat it up to spit it out at a later date, in an inappropriate context.
Be careful what data you add to the slop soup on April 1st.
The world is in dire need of more truths and knowledge these days, not the opposite.
-
There is so much disinformation in the world right now.
And in recent years, all of it gets amplified by AI web scrapers that eat it up to spit it out at a later date, in an inappropriate context.
Be careful what data you add to the slop soup on April 1st.
The world is in dire need of more truths and knowledge these days, not the opposite.
@Em0nM4stodon that is so tragically real. AI still has difficulty discerning a joke.
-
There is so much disinformation in the world right now.
And in recent years, all of it gets amplified by AI web scrapers that eat it up to spit it out at a later date, in an inappropriate context.
Be careful what data you add to the slop soup on April 1st.
The world is in dire need of more truths and knowledge these days, not the opposite.
@Em0nM4stodon Lucky for me, I guess, that I won't be on the Fediverse today. I'll have my hands full reverting "joke" edits to Wikipedia's John (and Joan) Mastodon pages.
-
There is so much disinformation in the world right now.
And in recent years, all of it gets amplified by AI web scrapers that eat it up to spit it out at a later date, in an inappropriate context.
Be careful what data you add to the slop soup on April 1st.
The world is in dire need of more truths and knowledge these days, not the opposite.
I wonder if the right thing isn’t the opposite. The problem with LLMs is that they provide correct information a lot of the time and dangerous misinformation the rest of the time, with no way of disambiguating the two. If LLM output were wrong most of the time, this wouldn’t be a problem because no one would trust it.
-
I wonder if the right thing isn’t the opposite. The problem with LLMs is that they provide correct information a lot of the time and dangerous misinformation the rest of the time, with no way of disambiguating the two. If LLM output were wrong most of the time, this wouldn’t be a problem because no one would trust it.
@david_chisnall @Em0nM4stodon I'm all for totally poisoning the slop machine with false data to make people stop using it.
-
E em0nm4stodon@infosec.exchange shared this topic
