Maximizing the minimum of $k$ submodular functions

Let $f(X)$ be a monotone submodular function from $2^{\{1, \ldots, n\}}$ to $\mathbb{N}$ and consider the problem of maximizing it subject to a cardinality constraint $|X| \leq m$. By using the greedy algorithm, it is possible to obtain a $1-1/e$ approximation (see, for example, these lecture notes).

But now suppose $f_1(X), \ldots, f_k(X)$ are monotone submodular functions and we want to maximize $\min(f_1(X), \ldots, f_k(X))$ subject to the same cardinality constraint. Is it still possible to obtain a constant factor approximation?

Krause, McMahan, Guestrin and Gupta, who call this problem Robust Submodular Observation Selection show that is is NP-hard to approximate this problem to within any constant.

StackExchange Computer Science Q#41315, answer score: 3