Is access date for online references and links worth providing?

Wikipedias often provide access date for their online references and links. Is it a worthwhile practice?

A related question that produced an interesting and worthwhile answer from Google Gemini: What are arguments against including access date?

Pro

 * Since online pages can change in time, access date serves to identify the version of the page.
 * The access date usually fails to identify the specific version, as per the argument "Access date could be useful in conjunction with Wayback Machine [...]".
 * Expanding on the above, stating access date for an online source is like stating a book edition for a printed book; it identifies the version of the text. And stating the book edition in a reference to a book seems uncontroversial.
 * Book editions are generally available in libraries, and sometimes in bookstores. By contrast, the text of a web page for an arbitrary date is generally not available anywhere (only a small selection of dates is in Wayback Machine). See also the argument "Access date could be useful in conjunction with Wayback Machine [...]".
 * Access date is required by some citation standards. (The specific standards should better be provided. ISO 690 possibly requires the access date for online publications in some of its national localizations. However, when asked "does ISO 690 require access date for online publications", Google Gemini denied that the date is strictly required; it stated that it is merely recommended. The actual situation is hard to verify since ISO 690 is behind a paywall.)
 * Access date is meaningful when no publication date is stated in the source and cannot be provided in the link/reference identification in the wiki (which is rarely the case, but is possible in principle).
 * A better policy is to refuse to cite web pages with no publication date on the page as unreliable since failing to identify the version properly. E.g. Britannica online does provide a date of last update in its articles, and even article history for some articles. Even newspaper articles, which are generally not truly reliable and independent, provide article dates.
 * That seems too stringent. E.g. Czech language web site Internetová jazyková příručka does not provide a date of last update in its dictionary entries yet it is generally reliable. (Better find English-language examples.) It seems better to allow provision of access date for such web sites.
 * Multiple arguments against apply, including that the version of the page text visible on that particular date cannot generally be retrieved from Mayback Machine given it covers only a small fraction of all possible dates of access.

Con

 * The access date is visual noise, making the reader experience worse.
 * The access date creates labor overhead.
 * The access date is usually practically redundant to publication date.
 * It is not redundant in so far as it identifies the actual version of the text, possibly different from the one present on the page at the date of publication. See also arguments for.
 * When the date of last change differs from the publication date, a serious publishing practice is to state something like "last updated on so-and-so", as practices e.g. by Britannica online.
 * Sure, and it is in case of both publication date and last update date present when the reader can quickly see that the page was accessed before the last update and that therefore the link/reference is no longer valid in a certain sense: the publisher decided to remove some of the information it previously had, possibly since they found it inaccurate, problematic, etc.
 * For the rare cases for which the access date is of any value, it can be retrieved from revision history of the wiki page.
 * The person entering the link could have accessed the page on a day earlier than the day of entry into the wiki.
 * Access date could be useful in conjunction with Wayback Machine. However, Wayback Machine does not archive the page daily or even weekly. Therefore, when one has an access date of, say, 20. January 2020, and the most recent version in Wayback Machine is of 1. January 2020, one cannot know the page has not changed between the two dates.
 * Comment: This could be addressed by requiring those entering access date to check with Wayback Machine and only use access dates for which archived versions exist. But this would create even more labor overhead.
 * Expanding on the above point in a somewhat different vein: "The content you get can depend on other factors than just the date. URLs often produce different content depending on your IP address, your user-agent, and cookies. It is even possible for a page to be written in different languages and automatically have it delivered in the language, which the user chose in their browser." Therefore, the attempt to uniquely identify the text via date fails anyway in principle even if that text is in fact available in Wayback Machine for that exact date.
 * The cases of page text depending on IP address, user-agent and cookies are probably much rarer than content changing because of update.
 * Comment: Identification sticklers could use CRC-32 of the page text or similar hash codes. There would have to be a standard way to determine what is included in that CRC-32; a web layout change impacts the HTML text and therefore also CRC-32. Almost no one or no one seems to be this sticklerish, though.
 * Access date is meaningful when the page is no longer available at its original location, to be entered after the page is no longer available.
 * If someone wants to preemptively enter the access date upon reference or link entry, it is unclear why they should not feel free to do so.
 * The access date is a visual noise and is better avoided as far as possible.
 * Preventing contributors from doing busiwork (work for work) redirects human resources to productive activities. And there are plenty of mundane valuable activities requiring no considerable talent or knowledge, e.g. systematically adding external links in Wiktionary. There are probably such valuable mundane activities also in Wikipedia.
 * Access date is not required in citation style guides APA and CMS; it was required in MLA but is no more. Inconclusive yet suggestive.
 * When one wants to add a piece of information traced to a reference already used multiple times on the page, one faces the trilemma: 1) either speculate that the referenced page has no changed since the last access and leave the access date unchanged, or 2) update the access date and be forced to check that all information traced still matches the text in the referenced page, or 3) create two references for the same page with different access dates. That is cumbersome and arguably not worth it.