Lawful Processing & Individual Rights, 30 min read

Sensitive data, children, marketing and profiling

In this lesson you will:

  • Apply the extra conditions that protect sensitive personal data
  • Handle children's data, including age assurance and parental consent
  • Run direct marketing that respects opt-in and opt-out rules
  • Recognise profiling and automated decisions, and the safeguards they need

Four areas cause a large share of privacy complaints and fines: sensitive data, children's data, direct marketing and automated decisions about people. Each adds rules on top of the basics from lessons 2.1 to 2.3.

1. Sensitive personal data

Processing sensitive data (lesson 1.1) usually needs two things: a normal legal basis and an additional condition, such as explicit consent, employment or social-security law, vital interests, substantial public interest, healthcare or legal claims. Several US states require opt-in consent for sensitive data, and California gives consumers a right to limit its use.

Practical safeguards:

  • collect it only where clearly necessary, and say so in the notice
  • separate it from ordinary data, with tighter access control and logging
  • run a risk assessment (lesson 3.3) before large-scale processing
  • watch for inferred sensitive data: an algorithm that infers pregnancy from shopping or health conditions from app usage is processing sensitive data even if nobody typed it in

2. Children's data

Children are less aware of risks and more open to manipulation, so laws protect them more strongly. Ages differ:

Law Who counts as a child for consent Key rules
India DPDP Under 18 Verifiable parental consent; no tracking, behavioural monitoring or targeted advertising directed at children; limited exemptions set by the Rules
EU GDPR Under 16 by default; countries may lower to 13 Parental consent for online services offered directly to children below the national age
US COPPA Under 13 Verifiable parental consent for online services directed at children or with actual knowledge
California Under 16 Opt-in to sale or sharing (parent for under-13s)
China PIPL Under 14 Treated as sensitive; parental consent and special rules
UK Age Appropriate Design Code Under 18 High-privacy defaults, no nudging, profiling off by default
Bar chart of the age below which parental consent or child protections apply: COPPA 13, PIPL 14, GDPR default 16 (13 to 16 by country), California 16, DPDP 18, UK design code 18.
Figure 2.10 The "age of a child" varies from 13 to 18. Global services often design to the highest age that applies.

Age assurance means working out whether a user is a child, in proportion to the risk: self-declaration for low-risk services, age estimation or verification through trusted credentials (for example, government-backed digital lockers) where risks are higher. Collecting more identity data than necessary to check age creates its own privacy risk.

Verifiable parental consent means taking reasonable steps to confirm that the consenting adult is really the parent or guardian, such as checking against identity details already held or a verified token.

🔐 Privacy lens

Design for children by default where you cannot reliably tell ages apart: high-privacy settings, no precise location, no public profiles, no behavioural advertising. It is usually cheaper than bolting protections on later.

3. Direct marketing

Marketing rules sit in privacy laws and in electronic-communications or consumer laws.

  • Email and SMS: the EU and UK require prior opt-in consent for individuals, with a "soft opt-in" exception for similar products sold to existing customers who were given a chance to refuse. The US CAN-SPAM Act allows email marketing with a working opt-out. India regulates calls and SMS through TRAI's commercial-communication rules and the national Do Not Disturb registry.
  • Phone calls: check do-not-call registries before calling.
  • Targeted online advertising: often relies on cookies and tracking (lesson 4.3); US state laws give an opt-out of "sharing" for cross-context behavioural advertising and require honouring Global Privacy Control signals in several states.
  • Always: honour opt-outs quickly, keep a suppression list so people are not re-added, and identify yourself.
Matrix of marketing channels against the default rule: email, SMS, calls, postal mail and online ads, showing opt-in or opt-out and the key safeguard for each.
Figure 2.11 Marketing rules by channel. When in doubt, ask for opt-in and make opting out effortless.

4. Profiling and automated decisions

Profiling is automated processing to evaluate or predict aspects of a person: interests, creditworthiness, reliability, health, location. It is common and often lawful, but it must be transparent and fair.

Solely automated decisions with legal or similarly significant effects, such as refusing a loan, rejecting a job application or setting an insurance price with no meaningful human involvement, carry extra rules. GDPR allows them only where necessary for a contract, authorised by law, or with explicit consent, and requires safeguards. Several US states give opt-outs from profiling that produces significant decisions, and PIPL requires transparency and an option to refuse personalised targeting.

Safeguards for significant automated decisions:

  1. tell people the decision is automated, the logic involved and the likely consequences
  2. offer meaningful human review by someone with authority to change the outcome
  3. let people express their view and contest the decision
  4. test models for accuracy and bias, and monitor outcomes over time
A loan application flowing through an automated scoring model. Low-impact outcomes proceed automatically; adverse decisions trigger notice, human review, the right to contest and bias monitoring.
Figure 2.12 Safeguards around a significant automated decision.

Worked example — a children's learning app

Area Decision
Audience Users aged 8 to 16 in India, the EU and the US
Age Design to under-18 rules everywhere; ask age at sign-up with neutral wording
Consent Parent account required; parent verified against a trusted credential, such as a government-backed digital locker token; child cannot change privacy settings alone
Data collected First name, class, progress; no photos, no precise location, no contacts
Profiling Learning-level recommendations only; no behavioural advertising; no tracking across other sites
Marketing Emails go to parents only, with opt-in; none to children
Notices Child-friendly first layer with pictures; full notice for parents
Review Risk assessment signed off before launch; reviewed when features change

Key takeaways

  • Sensitive data needs a legal basis plus an extra condition, tighter controls and attention to inferred data.
  • Child ages range from 13 to 18 across laws; DPDP sets 18 with verifiable parental consent and bans tracking and targeted ads directed at children.
  • Marketing rules vary by channel; opt-in, suppression lists and effortless opt-out are the safe baseline.
  • Significant automated decisions need transparency, human review, a way to contest and bias monitoring.

Self-check

  1. A shopping app infers that a customer is pregnant. Why does that matter?
  2. What does DPDP prohibit in relation to children, beyond requiring parental consent?
  3. Name two safeguards for a solely automated credit decision.