Elon Musk has made a big deal of X's free speech agenda. But it turns out — and who could have seen this coming — that approach to modern social media just doesn't make money.

Twitter was never profitable. But under Musk, the company known know as X has fallen even deeper into the red. In June, the company's ad sales were down 59% compared to the same time last year.

Now, X will test the effectiveness of a tool called sensitivity settings in wooing advertisers back to the platform.

In a blog post, the company said the tool would use machine learning to ensure that a brand's ads show up next to content that meet its unique sensitivity needs, which they also call a sensitivity threshold. It's a lot of big words that mean the same thing: brands don't want their ads showing up next to content that isn't good for PR.

A screenshot of the tool details the three levels of sensitivity:

• Relaxed sensitivity, which is not yet available, will show ads alongside some sensitive content to maximize reach but will exclude targeted hate speech and explicit sexual content

• Standard sensitivity is for brands with moderate sensitivity thresholds and will not show ads alongside targeted hate speech, explicit sexual content, gratuitous gore, excessive profanity.

• Conservative sensitivity is a strict sensitivity threshold that will not show ads next to targeted hate speech, sexual content, gratuitous gore, excessive profanity, obscenity, spam, and drugs.

The blog notes that Standard sensitivity is selected by default.

The blog also works overtime to assure advertisers that spend their money on X is a good idea. In addition to an endorsement from an SVP at Mondelez, the post offers up this wild claim: over the past 9 months, X has delivered a decade’s worth of innovation focused on creating an engaging and healthy environment where everyone, including advertisers, can connect safely.

But if X is so safe, why is it struggling to retain advertisers? I guess X's ads team can tackle that quandary next.