Good reframe. The question isn't whether the model is intelligent — it's what we let it decide.
Most of the failures I see aren't model failures. Someone handed a system decision rights nobody wrote down, and nobody owns. That's a design problem, not an AI problem.
Bounded permissions. Clear failure detection. A named human on anything that matters. Not exciting, but that's where it's actually won or lost.
It's an incredible technology, especially LLMs, but the integration with other tools, tht allows it to have a database, etc is what give us the potential that is giving us. This is not a tool that decides by itself do things, this is all pre program, and indeed the people that developed and set it up cannot avoid responsibilities.
Timely article as many AI fears are being tossed about. I totally agree this technology requires governance and legal guidelines put in place by governments.
It totally needs governance, I fear that let companies lead themselves on this is a mistake, the same way we need to also not only support our governments but also make them accountable. It’s a balance…
This is a great video that factual represents what really happened. This is not AI/LLM fault, but human problem, companies should be liable for their products.
Good reframe. The question isn't whether the model is intelligent — it's what we let it decide.
Most of the failures I see aren't model failures. Someone handed a system decision rights nobody wrote down, and nobody owns. That's a design problem, not an AI problem.
Bounded permissions. Clear failure detection. A named human on anything that matters. Not exciting, but that's where it's actually won or lost.
It's an incredible technology, especially LLMs, but the integration with other tools, tht allows it to have a database, etc is what give us the potential that is giving us. This is not a tool that decides by itself do things, this is all pre program, and indeed the people that developed and set it up cannot avoid responsibilities.
I think most of these news about bots going wild, it's just engineers not respecting procedures or bad implementation going wrong.
This is what happens when we try to use probabilistic models to perform deterministic tasks. It doesn't add up.
Timely article as many AI fears are being tossed about. I totally agree this technology requires governance and legal guidelines put in place by governments.
It totally needs governance, I fear that let companies lead themselves on this is a mistake, the same way we need to also not only support our governments but also make them accountable. It’s a balance…
This is a great video that factual represents what really happened. This is not AI/LLM fault, but human problem, companies should be liable for their products.
https://www.youtube.com/watch?v=yhHV5l1OA28