This kind of smells like a "gentleman's agreement" to me, if not implied threats of retaliation.
Speculation: The helicopter crews know that if they don't report traffic in sight as soon as it's called, they'll be slowed down and deprioritised by ATC and eventually held back until they either do report the traffic, or there's a substantial gap in arrivals - see the LH A380. That makes them unpopular with their passengers and/or superiors, so they are very
flexible with what 'in sight' means.
ATC likewise knows that if they push helicopter crews too hard on altitude busts, report anything involving a helicopter, or hold up either kind of traffic, they might get people breathing down their neck and certainly nothing good comes of it.
Calling traffic immediately and not enforcing separation too strictly allows both parties to 'get on with their jobs' while looking more-or-less by-the-book - until an incident like this happens.
I'm hopeful I'm wrong, but given there seems to be a long history of near misses and altitude busts this seems like the obvious conclusion.
No-one high-up wanted to hear about it or change anything, because no-one had died yet.
On a slightly different note, I'm curious whether anyone is familiar with the Hierarchy of Controls by NIOSH? It doesn't map 1:1 to aviation, but it codifies some things that are 'obvious' in hindsight:
Broadly speaking, some controls are more effective than others. Wherever possible, you should attempt to use more effective controls in place of less effective ones. More effective means not just that it reduces the risk the most, but also the most reliable over time and most resistant to having rules bent, being left broken, being ignored due to alarm fatigue, or 'normalisation of deviation'. Procedures that assume everything is working perfectly and everyone is 100% competent
will fail; see MCAS and a great number of other accidents.
Elimination is rarely possible but substitution (radar vs visual) and isolation (separate helicopters from other traffic) amongst other engineering controls are potentially more feasible, and much higher up the hierarchy than a glorified instruction not to crash (the very bottom of administrative). Engineering a problem out of existence is far superior to having a procedure to fix it in the QRH.
I list things like TCAS, GPWS, RSAs, and crash-proof seating as broadly being under PPE: they're nice to have and certainly worth pursuing, but unless there is no other alternative, they should never be your primary protection. Something has gone wrong if they get used.