KGI Applauds Introduction of State Legislative Provisions Aimed at Creating Better Algorithmic Feeds

As state legislatures begin their 2026 sessions across the country, the Knight-Georgetown Institute (KGI) is encouraged to see new legislative ground being broken on algorithmic feeds. In Alabama, State Representative Ben Robbins has introduced House Bill 171 (HB 171), which contains a novel and forward-looking approach to governing algorithmic feeds that includes provisions that go beyond “addictive feeds” bills that were enacted in New York and California last year.

HB 171 advances policy around online algorithms in two important respects, both of which we recommend in our Better Feeds report and model bill: transparency and user control.

Transparency: HB 171 would require social media platform operators to publicly disclose a description of each algorithmic system on its platform, each source of input to the algorithmic system on its platform, and the weights used (in aggregate) to determine the relative contributions of each input. To our knowledge, this is the strongest transparency measure aimed at making algorithmic parameters understandable that has been proposed anywhere in the country.

Requiring disclosure of the complete set of inputs and aggregated weights – numeric settings that determine how much each input factors into the algorithm – would allow the public and independent experts to understand whether and how platforms are seeking to maximize users’ attention by relying on clicks, likes, shares, and other engagement signals most heavily.

User control: HB 171 would require that users be allowed to unambiguously communicate their preferences about the types of items they want to see blocked or recommended on social media platforms, and to have those preferences honored. Too often, users choose to “hide” or “show more” of something they see on social media – and see no change in their experience. This provision aims to fix that.

HB 171 reflects a growing recognition among policymakers that AI-powered algorithmic recommender systems play a powerful role in shaping what people see, read, and watch online – and that these systems often prioritize short-term engagement over long-term user value and well-being. Lawmakers have the opportunity to incentivize better design of these systems on social media and across the online ecosystem including search, gaming, and online advertising.

Many policy responses have framed the issue as a false choice between engagement-optimized feeds and chronological or non-personalized feeds. As Better Feeds makes clear, there is an alternative: algorithms that put users first by prioritizing long-term user value over short-term attention maximization.

In addition to transparency and meaningful user controls, Better Feeds also highlights the need for user defaults and assessments of long-term impact. While HB 171 represents a significant step forward, policymakers have opportunities to build on this foundation by ensuring that platforms provide recommender system defaults designed to serve users’ long-term interests – especially for minors – and to continuously test and disclose the effects of algorithmic changes through extended “holdout” experiments that exempt a group of users from design changes for 12 months or more. The Better Feeds model bill lights the path towards these improvements.

“The transparency and user choice provisions in HB 171 reflect meaningful moves toward algorithmic governance that puts users’ interests front and center,” said Alissa Cooper, Executive Director of the Knight-Georgetown Institute. “By pairing strong transparency requirements with enhanced user control, Representative Robbins’ bill aligns with components of the Better Feeds framework that promote systems that deliver long-term value for users.”

Resources

Better Feeds: Algorithms that Put People First 
Model Legislation for Better Algorithmic Feeds