HMRC deve rivelare se ha utilizzato l’IA nelle decisioni di credito d’imposta, le regole del tribunale
https://www.ft.com/content/ca8e65f8-f2fc-4121-ab78-0d3f804cc01a
di 457655676
HMRC deve rivelare se ha utilizzato l’IA nelle decisioni di credito d’imposta, le regole del tribunale
https://www.ft.com/content/ca8e65f8-f2fc-4121-ab78-0d3f804cc01a
di 457655676
3 commenti
> The judgment, seen by the Financial Times, concluded that “the balance of the public interest lies in disclosing the information requested”
I agree with this. If someone at HMRC makes a decision that ruins someone’s life, because they consider their probability calculator infallible, then that’s dangerous. Especially if it’s an LLM (like ChatGPT/Claude) involved, which is quite possibly one of the stupidest things to do here, and hopefully that isn’t the case. If it’s an actual machine learning model used to detect fraud based on various parametric inputs, and based on well known methods, then that is fairly common especially in the financial world.
That would make the answer no because AI doesn’t exist. Now if they were ordered to disclose if they used an LLM alongside any other algorithms to automate decision making then that would be a different question entirely
All systems should be above reproach but surely they have other calculation software too? Hardly think it’s all being tallied up manually?
I’d expect the same level of checks and data security for any software tbh, or could be Post Office scandal take 2