Edit: I misunderstood the left/right requirement in my first pass. (Are Mulligans permissible with these teams?) Here's a version that's hopefully much closer to what you want:

my %players;
for (<DATA>) {
    chomp;
    my ($team) = m!^ (.+?) = (?<left>.+?) \s*/\s* (?<right>.+) $!x
        or die "Line `$_' doesn't match";

    for my $side (qw<left right>) {
        $players{$_}{$side} = $team for split m!\s*[,/]\s*!, $+{$side}
+;
    }
}

# Remove players only on one team
delete @players{ grep { keys %{$players{$_}} == 1 } keys %players };

local $" = ' AND ';
print "$_ IS IN @{$players{$_}}{qw<left right>}\n" for keys %players;
[download]

The logic is relatively straightforward. For each line, parse the team, left, and right strings or die trying, and then split the left and right chunks into left/right. We are then left with, for example, $players{Don} = { left => 'Team4', right => 'Team5' };. After that, we can delete any player who has only 1 key in their hash, as they are only on left or right, and not both.

I'll leave my original misguided answer below the readmore tag.

my %players;
for (<DATA>) {
    chomp;
    my ($team,$players) = split /=/;
    push @{$players{$_}}, $team for split m!\s*[,/]\s*!, $players;
}

delete @players{ grep { @{$players{$_}} == 1 } keys %players };
    
local $" = ' AND ';
print "$_ IS IN @{$players{$_}}\n" for keys %players;
[download]