Tumblr bans non-consensual creepshots and deepfake porn
Tumblr is to ban content that results in the "unwanted sexualisation or sexual harassment of others".
The social network says this includes non-consensual "creepshots" - sexually suggestive photos of clothed victims.
In addition, it said it covers "deepfake" pornography, in which computer-generated images of a subject's face are transplanted onto explicit photos or videos of someone else's body without permission.
The rule-change is due on 10 September.
Other additions to Tumblr's community guidelines include the outlawing of "gore" posted solely to shock, as well as other "content that encourages or incites violence, or glorifies acts of violence or the perpetrators".
The blogging platform will still rely on its users to flag offending content to moderators.
Yahoo had bought it for $1.1bn (£853m) the previous year promising not to "screw it up".
But Yahoo struggled to manage the platform, and ended up writing down most of its value.
In 2017, Yahoo itself was acquired by the US telecoms firm Verizon, which placed Tumblr within a new subsidiary called Oath.
While other social networks that allow pornography - including Twitter and Reddit - had introduced tougher rules against so-called creepshots and deepfakes, Tumblr's rules remained more vague until now.
This led it to be singled out by the news site Motherboard in March, which revealed that not only was the network hosting such imagery but also advice on how to create it.
"One Tumblr provides a detailed guide on how potential stalkers can get better quality photos, with sections on iPhones, Android devices, and DSLR cameras. A Google search of phrases in the guide suggest it is not available elsewhere online," it reported.
Tumblr says its staff will remain "fierce defenders of free expression" but added that the new rules were designed to protect that vision.
"Posting sexually explicit photos of people without their consent was never allowed on Tumblr, but with the invention of deepfakes and the proliferation of non-consensual creepshots, we are updating our community guidelines to more clearly address new technologies that can be used to humiliate and threaten other people," it said.
The move may help repair Tumblr's reputation, however it is likely that more will need to be done to reverse its decline.
One recent article about the platform likened using it to "hanging around a ghost town" despite the fact that Tumblr still claims more than 33 million posts are uploaded to its site a day.
The author - culture editor at The Outline, Jeremy Gordon - said Oath's current strategy seemed to be to make the experience of using Tumblr better for its remaining users rather than chasing new ones, at least at this point.
"It's hard to disagree with the value of discouraging hate speech, especially as we've seen how Twitter's total failure to address a similar problem has slowly shifted perception of the platform," he told the BBC.
"But as with all things like this it's in the execution - both consistency and the flexibility to understand not all situations are the same.
"At the very least, it's good to discourage any potential new users who'd only sign up to cause this kind of trouble."