Send one communication saying "your bot is clearly throwing up false positives; take action to investigate & remediate, ceasing all bot usage until implemented". Now they're aware / can't claim lack of knowledge. Give them a week or two to read the mail & turn the bot off, then any subsequent requests charge them for. If the fine's based on damages, charge them for all employees time taken to trawl through removing all spammy requests / for the cost of the team on your side who have to develop the automated filters.
> Send one communication saying "your bot is clearly throwing up false positives; take action to investigate & remediate, ceasing all bot usage until implemented". Now they're aware / can't claim lack of knowledge.
Yes, they can. The fact that someone made that claim about the reliability of the process doesn't mean that the person sending the notice knew that any particular notice was in fact false at the time they sent it, even notices made after the claim of unreliability.
It is possible that if it can be proven that the recipient read the complaint, and failed to investigate it because they believed it was likely to be correct (or if, even without such a notice, they knew of the unreliability of the process and failed to investigate the facts of particular notices and just blindly relied on the process), and that decision was motivated by a desire to avoid discovering that the information was false, then they might be considered to have constructive knowledge, but that's a far cry from "a notice was sent claiming that the process was unreliable, so any error resulting from that process automatically will be found to be 'knowing'."
