If you're replacing an old G83 inverter on an existing system then that's fine.
if your operating 'below the radar' as it were and were not planning on exporting significantly then they'll work fine, and if you already have a grid connected system back-feeding through the meter they aren't going to notice an occasional small extra backfeed.
On a small domestic system it probably isn't a big deal, but there was, a year or so back, that brief blackout in England somewhere which was started by a 500Mw fossil fuel plant shutting down unexpectedly, causing the local frequency to drop, then a new windfarm (also running about 500Mw) also shut down - and that caused the 'load shedding' blackouts to protect the grid. The issue with the windfarm was that it was still being built and they'd temporarily installed out-dated 2nd hand inverters to get some money rolling in. But the old G83 inverters shut down on grid under-frequency before the grid load-shedding threshold - which is the worst thing they could do of course, as the last thing you want to do with grid under-frequency is disconnect power sources. The wind farm got fined, although the real error is with the idiot who spec'd the G83 under-frequency disconnect Hz > the grid load-shedding Hz.
Technically though, if you're going to connect microgeneration to the grid with the potential to export up to 16A (3.68kw) then you ought to notify the network operator by filling in their form and they do ask for the inverter details and will expect it to be G98/99 as they don't want any more G83 blackout hazzard types connected; if you want to connect anything >16A potential export you need their permission, and if you fit a system that has the potential for >16A but has a facility to limit export to <=16A then the limiting part needs to be G100.
I think.
I don't actually do grid connect renewables and am not 'MCS' certified, so I am not certain i've got it right.
