Unfair Data Collection in Ad Targeting
Tags: Case Studies
Overview
Recent FTC enforcement actions against Mobilewalla and Gravy Analytics spotlight risks in creating audience segments derived from sensitive or location-based data without adequate notice or consent. These actions reinforce the applicability of Section 5 of the FTC Act to AI-assisted targeting and profiling.
Allegations Highlighted by FTC
- Creation of detailed audience segments such as religious groups, protest attendees, or frequent visitors to sensitive locations (e.g., reproductive health clinics).
- Lack of consumer notice and valid consent for use of sensitive and location data.
Key Principles Reinforced
- Transparency: Clear representations about data collection and use are expected; undisclosed sensitive segmentation can be deceptive.
- Fairness: Using inferred or sensitive characteristics for ad targeting without safeguards can constitute unfair practices.
- Consistency: Practices must align with consumer-facing disclosures and contractual commitments.
Implications for AI-Assisted Segmentation
- Even if sensitive attributes are not explicitly ingested, AI outputs may reveal or infer protected traits (e.g., emotional state, race, gender), attracting scrutiny.
- Implement governance to review segment labels, avoid sensitive inferences, and verify data sourcing (scraping, synthetic data, first-party sources).
Actionable Steps
- Conduct data protection assessments for profiling that may pose heightened risks.
- Build contract provisions addressing consent pathways, deletion, auditing, and restrictions on sensitive data usage.
- Establish ongoing monitoring and bias audits where appropriate to detect impermissible segmentation patterns.
Sources:- IAB_AI_Governance_and_Risk_Management_Playbook_August_2025.pdf
← Back to Knowledge Base