Engaging in a live debate requires, among other things, the ability to\neffectively rebut arguments claimed by your opponent. In particular, this\nrequires identifying these arguments. Here, we suggest doing so by\nautomatically mining claims from a corpus of news articles containing billions\nof sentences, and searching for them in a given speech. This raises the\nquestion of whether such claims indeed correspond to those made in spoken\nspeeches. To this end, we collected a large dataset of $400$ speeches in\nEnglish discussing $200$ controversial topics, mined claims for each topic, and\nasked annotators to identify the mined claims mentioned in each speech. Results\nshow that in the vast majority of speeches debaters indeed make use of such\nclaims. In addition, we present several baselines for the automatic detection\nof mined claims in speeches, forming the basis for future work. All collected\ndata is freely available for research.\n