Canada’s national library search service was hit by hundreds of suspicious requests in May and June, including attempts to probe its systems for security vulnerabilities.

Library and Archives Canada’s online collection search service was targeted by a series of attempted cyberattacks earlier this year, according to an analysis by cybersecurity research firm Transluce.

The activity occurred on May 28 and June 9, when roughly 900 requests were directed at the service. Many of the requests appeared to involve searches for digitized Canadian divorce records dating from 1905 to 1911.

Transluce identified 13 requests that contained attack payloads designed to probe the system for potential weaknesses rather than perform ordinary searches. Among the activity was an attempt to test the service’s record-identifier parameter for vulnerabilities.

The research firm did not definitively identify who was behind the activity. However, Transluce said some of the tactics resembled patterns it had previously associated with activity attributed to OpenAI.

Those similarities included the use of the web archive service Arquivo.pt, targeting obscure historical information and systematically probing websites for security weaknesses.

“We do not confidently attribute these attempts to OpenAI,” Transluce said, while noting that the activity exhibited tactics consistent with prior agent activity the firm had attributed to OpenAI during a similar period.

Canada’s national cybersecurity agency acknowledged the reported activity Tuesday and said it was assessing the situation with government partners.

“There is no indication that government systems have been compromised at this time,” the Canadian Centre for Cyber Security said.

The agency also noted that government websites routinely receive automated requests, including potentially malicious ones, as part of normal internet activity. Officials said reports of suspicious activity are nevertheless taken seriously.

The Cyber Centre said it uses layered security measures intended to identify and block malicious activity regardless of its source.

Transluce found no evidence that the attempted attacks were successful. The firm said there was no indication the library’s database processed the malicious input or provided additional information in response.

The incident highlights growing concerns surrounding automated and AI-assisted activity on publicly accessible websites, particularly as security researchers continue to examine how such systems interact with online databases and government services.