Facebook and Twitter revoke most of babysitter rating service Predictim's access to user data following report on how it scans social media to generate ratings
Facebook and Twitter are taking action against an automated babysitter-rating system reported on last week by The Washington Post …
Context & Ripple Effects
Predictim spent last week under scrutiny from The Washington Post: first for how it pressures babysitters to grant broad access to their social media so its AI can score years of online activity, then for the questionable character scores those evaluations produce for parents. The platform response was the missing piece — both Facebook and Twitter have now revoked most of the service's access to user data.
That matters because Predictim's entire product depends on that pipeline: without bulk social-media history there is no scoring model, and the revocation came within days of publication rather than through any regulatory process.
First-order effects
- Predictim loses the raw material for its ratings — with most Facebook and Twitter data access cut off, its character scores for new babysitter evaluations can no longer be generated at scale.
- Babysitters who were pressured into granting broad account access lose the immediate prospect of those evaluations reaching parents, while parents shopping for sitters lose a tool they had been marketed as due diligence.
Second-order effects
- Other startups building screening or scoring products on top of social APIs now face tighter developer vetting and a higher bar for demonstrating legitimate use before platforms grant broad data scopes.
- Background-check and childcare-marketplace competitors gain an opening to position human-vetted or consent-based verification against automated social-media scoring, which the reporting showed produces unreliable results.
Third-order effects
- The episode points toward platforms acting as de facto gatekeepers over who may algorithmically score individuals — enforcement by API revocation arriving faster than privacy or consumer-protection rules can, and setting precedent for how third-party profiling services are policed.
- If press-driven revocations become the norm, services built on scraping or broad-scope social data will need explicit consent architectures to survive, shifting the economics of people-scoring toward auditable, opt-in models.
The trend: Platforms are tightening control over third-party access to social data, cutting off automated people-scoring services when reporting exposes their methods rather than waiting for regulators to act.