HighWire at 25: David Worlock looks back - Highwire Press

"So we emerge into the world of AI. In fact of course it has been with us all along, changing its name from ‘Expert systems’ to ‘neural networks’ as a way of disguising the fact that it didn’t deliver to our wildest expectations. Now that it is finally beginning to deliver, it can shoulder many duties for scholarly research. All that peer-review reference checking and concept analysis, for example. But these are the foothills. It is when AI becomes the way that researchers read other people’s research that things really get interesting. Released from the time-consuming literature, researchers may be free to research and self-publish. But acts of self-publishing may simply be releasing formatted work into the network, or opening access to the network for a digital lab notebook. Imagine Jupyter notebooks of the future where colleagues and collaborators could see and annotate findings, or test reproducibility from the data available on board? As we move from Open Access to Open Science, overt acts of ‘publishing’ may become as rare as overt acts of textual reading. The minds of librarians (e.g. Hypergraph from Liberate Science) already lean in this direction...."