Right, thought experiment on how obsolescence works if it is tied to target players rather than the given player:
When a wonder becomes obsolete based on a tech that the current player has, that tech gets a hidden negative effect in that it removes one of the bonuses the player invested production in to achieve. As a result, the player will delay researching that tech until its benefits outweigh the new, negative effects. The player also has full control over when the wonder becomes obsolete.
When a wonder becomes obsolete based on a tech that another player has, that tech gets a hidden positive effect in that it removes one of the bonuses another player invested production in to achieve. As a result, the player is incentivized to research that tech even more. The player who invested production into the wonder does not have control over when the wonder becomes obsolete against any particular player.
As a result, while the current obsolescence mechanic works best when it applies to a tech that the player who built the wonder would want anyways, the proposed obsolescence mechanic that works based on the other players' tech levels works best when it applies to a tech that the players would not necessarily want without the added obsolescence bonus. In both cases, tech trees that contain more techs and do not have clear-cut, mutually exclusive ideal tech paths work best, since players can accommodate their shifted priorities better.
The problem is that the Great Wall in Civ5 follows none of these. Civ5's tech tree has roughly three "best" tech paths (two in multiplayer) that become mutually exclusive around Renaissance: one aims for lategame victory thanks to research labs through Plastics, one aims for a timing push with artillery through Dynamite, and the third one aims to quickly grab Archaeology for the extra tourism before heading into Plastics. Great Wall becomes obsolete through Dynamite, so in the current system, the builder will wish to avoid Dynamite, while the proposed system, the attackers will wish to prioritize Dynamite. This adds nothing to the game: if the builder builds Great Wall for its bonus, they are not looking to win with a timing push, so they would not prioritize Dynamite anyway and lose nothing by delaying it. In the proposed system, the players who suffer most from Great Wall, and therefore the ones looking to obsolete it as soon as possible, are the ones who are looking to end the game before Information Era, ie. they will want to win with a timing push; since Dynamite's artillery are so important to timing pushes, attackers would be prioritizing Dynamite anyway, so the fact that Great Wall's bonus is negated at dynamite does not change tech paths in any way.
In Civ5's current state, obsoleting the Great Wall at Dynamite does not add to the game the same way obsolescence did in previous Civ games, regardless of whether the obsolescence works based on the builder's tech level or the attacker's tech level. Even if Great Wall were to become obsolete at a tech like Industrialization, it would not work as well as obsolescence worked in previous Civ games: the fact that Civ5's tech tree has such clear-cut "best" tech paths mean that even if the Great Wall's builder would wish to avoid Industrialization, there is no real alternative techs that they could go for, they would just be forced to take the hit for those fast Research Labs. Likewise, the fact that artillery are so out of the way of other techs mean that if you wanted to obsolete another player's Great Wall, you would have to sacrifice so much time to do so that it's probably not worth it (remember, you're trying to win via a timing push with artillery).