ERIC, a clearinghouse and database for education research articles, has this notice linked from their home page:
Dear ERIC Community,
In early August we discovered that sensitive personally identifiable information appeared in some full text documents contained in the ERIC collection. Specifically, social security numbers and other highly sensitive information were found in multiple documents and in a way that could not easily be isolated. For that reason, we had to temporarily disable access to many full text documents.
Although these documents had been publicly available in microfiche for many years, the advent of Internet search engines has made it easier to find this information. Our number one concern is to ensure that any full-text documents we provide do not violate any individual’s privacy. We believe that if any of us were to have our privacy compromised by an ERIC document, we would want the same consideration.
We are seeking to restore access to documents as soon as possible. In order to restore access to ERIC, we have to check every document to see if it contains personally identifiable information. Due to the quality of many of the documents, a large portion of the search has to be done by hand. This is a large undertaking and we are in the process of hiring a team to help restore access in a fast and responsive manner. We hope to get this team in place by late September and releasing large numbers of ERIC documents by the end of October. We will continue to release documents after that point on a rolling basis.
To minimize the burden on our users, we will prioritize searching the documents that users request. If you would like to request a PDF to be returned online, please email [email protected] with the record number (such as ED263102). Documents will be returned on a rolling basis and may take several weeks, but we are working as fast as possible.
We are sorry for the inconvenience and want to thank you for bearing with us through this unexpected delay.
The ERIC Team
Given how old many of the articles and submissions in their database are – and many are not just reprints of conference submissions or publications – this is not a surprising situation at all. And it will be a mammoth job for them to go through every article to ensure that all non-necessary PII are redacted.