Why scraping profiles is a risk for staffing agencies
A profile anyone can see is not free to collect. What platform terms, Canada’s privacy commissioners, a US court case and the anti-spam law say about scraping, and safer ways to find candidates.
6 minute readSourcing and outreach
Scraping candidate profiles exposes a staffing agency on four fronts. It breaks the terms of the platforms your recruiters work on, which puts their accounts at risk. Canada’s privacy commissioners have rejected the idea that a profile anyone can see is free to collect. The email addresses it produces do not come with consent under the anti-spam law. And in the best-known court fight over scraping a professional network, the company that did it agreed to a US$500,000 judgment, to delete its data and never to scrape the site again.
What counts as scraping
Scraping is the automated copying of information out of web pages. In recruiting it takes a few common forms: programs that crawl profile pages, browser extensions that save the contents of each profile a recruiter opens, tools that find contact details to enrich a list, and vendors that sell profiles collected in one of those ways. The method matters less than the pattern: information copied in bulk, by software, without the people concerned or the platform agreeing to it.
The platform’s terms and your recruiters’ accounts
LinkedIn’s User Agreement, in the version effective November 3, 2025, is explicit. In section 8.2, members agree not to develop, support or use software, scripts, robots, crawlers, browser plugins or add-ons to scrape or copy the service, including profiles. They agree not to use bots or other unauthorized automated methods to access the service, add or download contacts, or send messages. And they agree not to copy, use or distribute information obtained from the service, directly or through third parties such as search tools, data aggregators or brokers, without the consent of the content owner. In section 3.4 of the same agreement, LinkedIn reserves the right to restrict, suspend or terminate your account if you breach the agreement.
For an agency, the exposure lands on individual people. An extension running in a recruiter’s browser runs on that recruiter’s account, so the account at risk is the one your team uses every day to search, message and keep in touch with candidates.
What happened to hiQ
hiQ Labs, a data analytics company, collected data from public LinkedIn profiles, and its dispute with LinkedIn went to the US Supreme Court and back. hiQ won a significant point along the way: the Ninth Circuit found it had raised serious questions that scraping public profile data was lawful under the US federal computer-access statute. It was then facing a ruling that it had breached LinkedIn’s User Agreement by scraping and by using fake accounts.
On December 6, 2022, the parties filed a stipulation and proposed consent judgment. It provided for a US$500,000 judgment against hiQ, a permanent bar on scraping LinkedIn, whether logged in or not, and the deletion of hiQ’s LinkedIn member data and of the code built to collect it. Those terms were agreed between the parties rather than found by the court, and they set no precedent. The lesson for an agency is a practical one: winning the argument that public data is fair game under one statute did not protect hiQ from the contract it had accepted.
Canadian privacy law: visible is not the same as available
PIPEDA lets an organization collect personal information without consent when the information is publicly available, but only if it is also specified in the regulations (section 7(1)(d)). The regulations name a short list: telephone directories, professional or business directories and listings, registries kept under a statute with a right of public access, records of courts and tribunals, and publications such as magazines, books and newspapers where the person provided the information. Most of those come with a further condition, that the use relates directly to the reason the information appears there. A social or professional network profile is not on the list.
The commissioners said so in the Clearview AI investigation, reported on February 2, 2021. The Privacy Commissioner of Canada, Quebec’s Commission d’accès à l’information and the commissioners of British Columbia and Alberta found that information from social media or professional profiles, collected from public websites and then used for an unrelated purpose, does not fall under PIPEDA’s publicly available exception. Collecting it requires consent, and then only for a purpose a reasonable person would consider appropriate in the circumstances, which is the test in section 5(3). They added that scraping in breach of a platform’s terms would be a further factor in finding a purpose inappropriate.
Clearview was a facial recognition case, far from recruiting, but the reasoning about public profiles is general. In October 2024, the Privacy Commissioner of Canada and 15 other privacy authorities issued a concluding joint statement on data scraping. It repeats that publicly accessible personal information is still protected by privacy law in most jurisdictions, says that a platform’s contract permitting scraping cannot by itself make the scraping lawful, and tells commercial data scrapers to take note.
Anti-spam law: an address you found is not consent
Scraped contact details usually end up in outreach, and that brings in Canada’s Anti-Spam Legislation. One route to implied consent is a conspicuously published address: the person published the address, or had it published, without a statement refusing unsolicited commercial messages, and your message is relevant to their business role. In the CompuFinder case, the Federal Court of Appeal said that this rule does not permit mining addresses from third-party directory websites, that the sender has to prove every element, and that a job title alone does not necessarily establish relevance.
An address a tool pulled from a profile or a directory rarely carries consent you can prove.
Buying the data does not move the risk
A dataset bought from a vendor raises the same questions in a different order: where did the vendor get it, and did the people agree? LinkedIn’s terms reach information obtained through data aggregators and brokers. PIPEDA’s consent principle expects the organization that hands over a list to obtain consent before disclosing it, which is a promise you would be relying on without being able to check.
There is also the conversation with the candidate. When someone asks what you hold about them, PIPEDA encourages you to tell them where it came from. “A tool copied it from your profile” is a hard answer to give a person you hope to place.
What to use instead
The safer channels are the ones where the person applied, agreed, or a platform licenses the data to you:
- your careers page and application form, with a privacy notice the candidate sees before sending anything;
- job boards, used through the board’s own tools and terms;
- referrals, where the referred person applies themselves;
- candidates who upload their own profile export;
- platform partner programs that license member data under contract.
Recruiters can keep using their own platform accounts by hand, within the platform’s rules. What changes is the bulk copying.
For the database you already have, record where each candidate came from and on what basis you hold the record. If a list came from a tool or vendor that copied it from a platform, stop using it for outreach until you can say how it was collected.
In practice
TalentInitiator has no scraping connector. Every candidate record carries its source and one of three legal bases: the person applied, the person consented, or a contract licenses the data. Our responsible sourcing page sets out the policy and the channels we use instead.
How to read this post
This post explains the rules in general terms, as they stood on the date above. It is not legal advice for your organization.
Sources
- PIPEDA Findings #2021-001 (Clearview AI) Decided February 2, 2021
- Personal Information Protection and Electronic Documents Act, s. 7(1)(d) (publicly available information)
- LinkedIn User Agreement, s. 8.2 Version of November 3, 2025
- LinkedIn User Agreement, s. 3.4 (restricting or ending an account) Version of November 3, 2025
- hiQ Labs, Inc. v. LinkedIn Corp. (consent judgment) Decided December 8, 2022
- Joint statement of privacy regulators on automated data collection from websites Published October 28, 2024
- Personal Information Protection and Electronic Documents Act, s. 5(3) (appropriate purposes)
- CompuFinder Management Inc. v. Canada (Attorney General)
- Personal Information Protection and Electronic Documents Act, Schedule 1, principle 4.9 (individual access and correction)
- Canada's Anti-Spam Legislation, s. 10(9) to (14) (implied consent, existing relationships) In force since July 1, 2014
- Personal Information Protection and Electronic Documents Act, Schedule 1, principle 4.3 (consent)