Common Sense Media rates AI products on a separate risk rubric

The design

Common Sense Media, best known for age-based content ratings, runs a separate AI Ratings program that scores generative AI products against a published methodology, distinct from the organization's long-running Privacy Program score for apps generally. According to that methodology, as archived 18 September 2024, an “AI Review Team” collects a “participatory disclosure” from a product's maker, adapted from AI-transparency formats such as model cards and datasheets for datasets, then conducts a “sociotechnical assessment” built on Santa Clara University's Markkula Center for Applied Ethics framework. That assessment covers a product overview, a “societal landscape assessment” of who benefits and who is affected, and an “ethical risk sweeping” exercise the methodology says is conducted by a team rather than one reviewer, to avoid a single person's blind spots driving the outcome.

What the evidence says

Common Sense Media's rating of Snapchat's My AI, dated by the site as last updated 19 October 2023, applies this methodology to a named companion-adjacent product and states plainly “we see more downsides to My AI than benefits.” The same page lists a separate “Privacy Rating” of 64%, drawn from the organization's distinct Privacy Program evaluation and linked out to that program's own page: two different scores, from two different rubrics, shown side by side. The review also credits Snapchat with two specific safety mitigations: a “time-out” triggered when the system detects abuse, and surfacing a safety-resources page in response to mental-health-related messages.

What it asks of people

A reader using this rating has to track which number answers which question: the AI Ratings score reflects the ethical-risk methodology described above, while the Privacy Rating reflects a separate data-practices evaluation covered elsewhere in this archive. Treating either figure as a single composite grade would misread how Common Sense Media itself says the two programs work.

Privacy and safeguards

The My AI review states “we spoke with Snap and they are working to investigate and address our findings,” which documents engagement with the company but not a confirmed outcome; neither this entry nor the cited pages establish what, if anything, changed as a result. The methodology page describes the disclosure and assessment process in detail but does not claim Common Sense Media independently tests a product's behavior beyond what the maker discloses and what reviewers observe using it.

• Has Snapchat's My AI been re-reviewed under this methodology since October 2023?
• Does the AI Ratings program disclose which specific claims in a company's disclosure it was able to verify directly?
• How does a team-based review process resolve disagreement among its own members before publishing a score?

Two numbers on one page can measure two different things; Common Sense Media's own methodology and its My AI review are worth reading together precisely because they do not collapse ethical-risk assessment and privacy practice into a single verdict.