The AI Alignment Forum

MARCH 25, 2025

In principle, we could even allow for an end-to-end game where competitors submit either a red team attack policy or a blue team protocol and then we see how the dynamics go. These settings can be interesting to study even if the blue team has access to a basically perfect but expensive oracle of whether the AI's actions are problematic.

Work 52

Work Evaluation Train Training 52

The Verge

MAY 4, 2021

You think about— Oracle had one, SAP had one, ATG, and Hybris, and IBM WebSphere, they were very expensive and very difficult to operate. Shopify obviously has a content policy. Well, the reason we have the accessible use policy is so that we have very clear lines. That contravened and went against the AUP policy.

Amazon 80

Amazon Network Business Channel 80

The Verge

AUGUST 20, 2020

Facebook also took down thousands of accounts, pages and groups as part of what they called a “policy expansion,” seeking to limit violent rhetoric tied to QAnon, political militias and protest groups like antifa. But coded messages are harder to discern, particularly by policy teams that have not invested in unscrambling them.

Myanmar 139

Myanmar Facebook New York Taiwan 139

The Verge

SEPTEMBER 29, 2020

Here’s Olivia Solon at NBC News : A group of about 25 experts from academia, civil rights, politics and journalism announced Friday that they have formed a group to analyze and critique Facebook’s content moderation decisions, policies and other platform issues in the run-up to the presidential election and beyond. Alex Thompson / Politico ).

Facebook 68

Facebook New York Policy YouTube 68

The Verge

AUGUST 28, 2020

That said, a Microsoft-Walmart pairing still seems preferable to me (and reportedly to Zhang) than the other big player in the hunt: Larry Ellison’s Oracle, which has spent the entire Trump Administration flattering the president’s ego and may now be poised to reap the rewards. TikTok has surprised me with its cleverness before.

Oracle 108

Oracle Facebook Hurricane Social Media 108

The Verge

SEPTEMBER 18, 2020

Oracle will also own a minority stake that will be less than 20% of the new global TikTok, two of the people said. Trending up : Facebook announced new policies to limit the spread of groups that promote violence and share health advice. You can read the full policy change here. It will no longer recommend these groups to users.

Fun 72

Fun Facebook Oracle Social Network 72

The Verge

SEPTEMBER 9, 2020

Goodman, a law professor at Rutgers University specializing in information policy, approaches the problem from another angle. Under a second, platforms would disclose data on what content is being promoted and to whom, on the process and policies of content moderation, and on advertising practices. Barrett writes (PDF): Ellen P.

New York 59

New York Twitter United States Campaign 59

The Verge

SEPTEMBER 25, 2020

As on Twitter, the app opens to a central scrolling feed of updates from people and topics that you follow. This is a policy written by people who have argued online, and who want to create a place where those arguments are productive. But as I like to say, policy is what you enforce. That all sounds true enough.

Social Network 91

Social Network Twitter Network Social 91

The Verge

AUGUST 19, 2020

Enforcement of those misinformation policies has sometimes lagged behind the companies’ public statements, though. In 2020, this is what successful content policy looks like: you can’t prevent every bad thing from ever being uploaded, but you can identify it quickly and take effective action. Jennifer Elias / CNBC). Industry. ?

Video 105

Video Platform Oracle Facebook 105