An important challenge in underlay dynamic spectrum access (DSA) is how to establish an interference limit for the primary network (PN) and how cognitive radios (CRs) in the secondary network (SN) become aware of their created interference on the PN, especially when there is no exchange of information between the primary and the secondary networks. This challenge is addressed in this paper by present- ing a fully autonomous and distributed underlay DSA scheme where each CR operates based on predicting its transmission effect on the PN. The scheme is based on a cognitive engine with an artificial neural network that predicts, without exchanging information between the networks, the adaptive modulation and coding configuration for the primary link nearest to a transmitting CR. By managing the tradeoff between the effect of the SN on the PN and the achievable throughput at the SN, the presented technique maintains the change in the PN relative average throughput within a prescribed maximum value, while also finding transmit settings for the CRs that result in throughput as large as allowed by the PN interference limit. Moreover, the proposed technique increases the CRs transmission opportunities compared to a scheme that can only estimate the modulation scheme.