AOL struggles mightily with spam email. It has millions of subscribers who receive hundreds of millions of spam emails per week. At enormous cost, AOL (and every other provider of email service) has to try to filter out this dreck.
A core problem is that the cost of sending email is very low. It's easy to buy a server that can pump out millions of emails per day, and the service providers have to receive that email and deliver it to their users. It uses a significant percentage of the available bandwidth on the Internet.
AOL has proposed charging a small fee to deliver email from bulk senders. This approach of charging for email has been around for a long time. If it cost, as an example, 1/100 of a cent to send an email, it would be a barely noticeable fee for most people, who send out just a few a day--if you sent 20 emails a day, it would cost you six cents a month for email. But if you were a spammer sending out a million emails, it would cost you $1000/million emails. All of sudden, sending spam costs real money. Spam would stop overnight.
But the problem is a bit more complex. Many nonprofit and civic organizations also send out lots of email, and AOL's current plan would not discriminate--senders of large amounts of email would have to pay. In return for payment, AOL would guarantee that an organization's email would get delivered to the recipient's mailbox (i.e. not get tossed in the spam bin).
But charging some for email delivery and not others would create a powerful incentive for AOL to put most resources into delivering the paid mail and correspondingly less into unpaid mail. So if you are sending a single (free) email to an AOL subscriber, it would get treated differently.
This is not a simple problem, and there are no obvious and simple solutions. But creating multiple classes of service on the Internet is likely to cause as many problems as it solves. It would be a difficult transition to go to fee-based email, but doing it across the board so that every email continues to get the same level of treatment would be better than AOL's proposal.