In general, the AMI consortium embargoes new annotations and new results of automatic processing for six months after they become available for consortium use, and then releases those which they think could be of use to the wider community here.

Annotation release v1.1, which came out in June 2006, contains some hand annotation for topic segmentation, abstractive summaries, named entities, individual actions (heads, hands, and movement around the room), and focus of attention. We expect to release all annotations of these types that are being produced by the consortium by January 2007; this is expected to include complete discourse annotations (topic segmentation, abstractive summaries, named entities, dialogue acts, and extractive summaries) for all scenario meetings, but not for non-scenario ones.

ASR output using manual segmentation and based on close-talking microphones will be released under the same terms as the rest of the data at the end of May 2007. Please contact us if you have a specific use for the data, such as an international evaluation, that would require a release before this, since we are happy to consider releasing it earlier under more limited terms.

We expect to release fully automatic ASR, still based on close-talking microphones, some time in the second half of 2007.

Note that because we use Share-Alike licensing terms, new annotations or new results of automatic processing may be produced and released under the Share-Alike license at any time. Although this process is outside the consortium's control, we will try to incorporate new annotations of which we are aware into our release structure, and encourage anyone who intends to produce them to contact us.