This should be exploited by challenging the mutuality of the agreement. Unless there is an understanding by both parties about the basic features and requirements - a "meeting of the minds" - then there isn't a contract.
Currently the "yes, I read that" buttons are used as an indication of having read and understood the contract, but as you said, actually reading all that legalese would take a long time. It may have been hard to prove otherwise in the past, but today we have another option.
We now have algorithms that estimate a given text's reading level. Using that kind of technique we can algorithmically estimate how long it should take for someone to read a document, from which a conservative estimate of the minimum reading time can be derived. It would be short enough that practically everybody exceeds the estimated time. This needs to be purely mechanical.
With a minimum reading time established, every contract should be nullified unless each party was given and used at least that much time to read the document. That is, allowing anybody to "sign" an EULA before $MINIMUM minutes have elapsed should be prima facie evidence that no contact exists.
"But until terms of service are both standardised and codified with users' interests in mind"
But what about sites with a different revenue model? Say an image hosting site meant for personal photos, and an image sharing site meant as a platform for artists to sell their work. You would not want the former allowing any sort of copyright transfer, but the latter cannot function without it(at least without the ability for the site to take payment and provide a license). Im not sure how we could standardize everything. Especially if consumers are not willing to pay a subscription.
"In particular, ordinary transactional terms are limited by:"
1. But this would be an ongoing relationship, assuming they continue to use the service.
2. Again, subscription fees would be needed to sustain the business. Otherwise advertising is the only way. And there's an old joke in advertising: half of all advertising budgets are wasted, the trick is figuring out which half. I do see your point about a drivers license number not being needed for a long period of time.
It's not clear to me how some breadth of interests couldn't be addressed. The usual T&C generally address limits on liability of the site's owners, occasionally try to impose binding arbitration or limits on class action suits (among my complaints against the so-called "Kinder, gentler Reddit", Imzy), jurisdiction, reverse-engineering clauses, etc.
Allowing users to specify licenses for submitted works would address much of your concerns. A standard set of merchandise clauses, including, say, escrow, liability, and chargeback terms, might be among the boilerplate additions to a standard contract which might be made.
But the point is to make the contracts themselves standard and modular. There might be a base services contract, a base merchant contract, and a base rights-for-sale contract, but not infinite variations on each. Also limits on what sites or users might carve out as grantable or transferable rights.
My point in noting that ordinary commerce is limited to a single sale transaction is just that: that these are simple transactions and hence the associated legal binding is also simple. Ongoing relationships are inherently more complex.
There are alternatives to advertising and subscriptions, including non-market constructs.
In the example I posed, the value of the drivers license as a hire surity is that the hirer is quitely likely to return for it. The disadvantage, today, is that the license has not only the attributes of "valuable to the owner", but "hive of data which can be used to draw additional relations".
Some years ago I discovered that the purchase of certain over-the-counter medications required, by store policy though not by local law, presentation of a drivers license. I held up my license for the clerk to visually examine. He tried to take it from my hand, which I refused. He wouldn't close the sale without scanning the card. I walked off without paying and without product.
I've been insisting on respecting my privacy rights for some time, and am not above forgoing business, taking my business elsewhere, or making others pointedly uncomfortable for asking questions I won't answer. Sadly, I am an exception.
This problem was solved in an earlier age of commerce through a Uniform Commercial Code (throughout most of the US), or equivalent statutory or case law in other domains. Essentially, contracts were reduced to a common set of standard components. Exceptions might be allowed for specific cases, including unilateral "contracts of adhesion", but these too were generally limited.
In particular, ordinary transactional terms are limited by:
1. The scope and extent of the transaction. The term is for a single purchase or transaction, not an ongoing "relationship", with few exceptions (utilities, rental agreements, subscriptions).
2. There's a limit to the data exchanged. In general, the minimum amount of data required for a transaction is what's provided. Even where personal information was taken down, it was recorded on and remained on paper forms, rarely being converted to electronic form. This is no longer the case, where, say, a license number (or a license held as surity) for, say, an hourly boat or bicycle hire might be scanned electronically, OCRd, entered into a database, and matched with other records.
3. As noted above, information one organisation gathered on you was only rarely shared with others. This is no longer the case. As my awareness of such practices has spread, I've become vastlty less interested in transactions in which I'm aware my information is being exchanged: magazine subscriptions, credit or debit card purchases, anything with an email or postal code, etc. For a good decade or more, I refused to sign electronic signature pads. I still generally balk at this.
But until terms of service are both standardised and codified with users' interests in mind, the present situation will only get worse.