Provenance is a concept in fine art collecting where a piece’s ownership is traced as closely as possible to establish a line from the artist to the current owner. Fine art collecting takes advantage of the physicality of the objects as much and whenever possible to corroborate the story the provenance tells.
Does the frame carry the marks of the builder that match other marks on other frames?
Does the sculpture show signs of the flood reported in that building when the piece was said to be there?
Does the impasto texture of the paint match that of other works by the artist?
When it comes to tracing the origin and authority of information, it has gotten a lot trickier. Much of the information presented on the internet has no additional dimensions or context which adds to the difficulty of tracing its origin. You can’t simply turn the blog post over to see if it is signed by the author and if that signature matches in form and substance to other signatures confirmed to be the author’s. If you are reading a post online and have no direct or physical connection to the producer or writer, it is quite hard to know without a doubt that the information is authentic.
The dawn of generative AI in 2023, which has been used to fabricate content of all forms, has once again muddied the waters of digital information provenance. Millions of AI generated videos are passed off as legitimate content. By simply browsing comments on TikTok and other short form video content platforms it is plain that the relevance of the Voight-Kampff Test or Turing Test has surely passed; there is no silver bullet or sure way for humans to be able to tell whether the content they are consuming was originated by a dedicated human laboriously typing each word out, an efficient human prompting a generative AI tool to write a piece and then reviewing the output for appropriateness, or an AI agent given permission to prompt a generative AI tool to mass-produce words in a constant string with little oversight.
Of course, in this environment what it means to create a work is complicated by different ways to view what the activity of creation fundamentally consists of, and which activities deserve automatic copyright protection, and which do not.
Google’s Gemini and Nano Banana image generators have taken a step in the right direction with SynthID, which Google DeepMind’s page explains is a, “watermarking tool, designed specifically for AI-generated content. It empowers users to identify AI-generated (or altered) content, helping to foster transparency and trust in generative AI.” If an image was edited or generated with Google Gemini, that activity can be “seen” by Google’s SynthID detector and reported to a user asking about an image’s provenance. This is a positive step towards a more legible and understandable version of the cyclone of information that is the internet.
The leadership of the library, too, considers consequences and makes decisions about how the library presents itself. The Baldwinsville Public Library has opted for increased authority. During the library’s recent website overhaul, whether the library’s web address, www.bville.lib.ny.us, would remain the same; could it be time to switch to a less cumbersome and more recognizable or more easily communicated string of letters and dots? This domain brings with it some added authentication guaranteed by the United States Federal Government. Bob Loftus, the Baldwinsville Public Library’s technology administrator and Systems Librarian, explains that, “in 1993, the Internet Engineering Task Force (IETF) created special rules for public library domains. Unlike .org or .com addresses that may be purchased from online registrars, the ‘.lib.ny.us’ domain requires an organization to verify themselves to the registrar, to prevent online impersonation of public organizations. That registrar is a Federal Government contractor responsible for verifying the integrity of public organization domains.
More recently, many libraries have switched to using .org domains that are sometimes easier for patrons to remember and type, but it’s much easier for hackers to spoof those domains or create ‘intentional misspellings’ of URLs (a common tactic of internet spammers and hackers). If we did not use the verified domain, a spammer might be able to send out mass phishing e-mails to library patrons with a URL like ‘bplibrary.org,’ and fool library patrons into providing sensitive information like credit card numbers to a malicious third party. The verified URL is just one little step that helps to protect library and patron communications.”
In this time of increasing uncertainty about where the information you consume comes from, the library leadership thought it prudent to choose the option that incorporates the most integrity for the provenance of the information it shares.