My own view differs from the mainstream view. I would say that gravitons have already been detected. I'm in favor of using a rigorous definition of "detection" that is not based on arbitrary rules that we've invented. Any argument about "detection" should be discussed from the laws of physics themselves. I therefore reject the notion that "detection" should always involve some traditional experiment, which in this case should involve some scattering experiment. You can't demonstrate, starting from the laws of physics, that this is the only way to establish the fact that gravitons indeed exist.

The existence of gravitons became a certainty when the existence of gravitational waves became a certainty. And that happened not when LIGO detected the signal from the black hole merger, but much earlier when tests of general relativity were performed back in the 1960s which confirmed its validity in at least the weak field limit.

This proves the existence of gravitons because the moment you have established the validity of General Relativity in the weak field limit, the fact that perturbations in the metric will propagate as gravitational waves is established. Quantum mechanics then implies the existence of gravitons. There is no loophole here, e.g. one cannot argue on the basis that perhaps quantum mechanics might fail.

While quantum mechanics might indeed fail in some regime, it is certainly valid in the low energy regime appropriate for describing low energy gravitational waves. Also note that non-renormalizability of quantum gravity is irrelevant, all that means is that higher order perturbative corrections are not independent of the details of physics at the smallest length scales. To calculate higher and higher order corrections you need to invoke more and more such details, which mans that lacking any knowledge of that, you can't do perturbative calculations in a systematic way. Renormalizable theories, in contrast depend only on a fixed, finite number of such details, which can be eliminated in favor of a few experimentally accessible physical quantities.

But none of this has any bearing on the existence of gravitons. At most one could argue whether or not the graviton would be a fundamental particle. But then neither would any traditional scattering experiment shed light on such an issue.

My own view differs from the mainstream view. I would say that gravitons have already been detected as I'm in favor of using a rigorous definition of "detection" that is not based on arbitrary rules that we've invented. Any argument about "detection" should be discussed from the laws of physics themselves. I therefore reject the notion that "detection" should always involve some traditional experiment, which in this case should involve some scattering experiment. You can't demonstrate, starting from the laws of physics, that this is the only way to establish the fact that gravitons indeed exist.

The existence of gravitons became a certainty when the existence of gravitational waves became a certainty. And that happened not when LIGO detected the signal from the black hole merger, but much earlier when tests of general relativity were performed back in the 1960s which confirmed its validity in at least the weak field limit.

This proves the existence of gravitons because the moment you have established the validity of General Relativity in the weak field limit, the fact that perturbations in the metric will propagate as gravitational waves is established. Quantum mechanics then implies the existence of gravitons. Now, as Jerry Schirmer points out in the comments, GR may break down before one reaches the regime where we could detect single graviton processes. But gravitons then still exist as quasi-particles at lower energies, and single graviton processes do occur at those lower energies, it's just that we won't be able to see it there due to technical issues.