On 11 October 2026, Home Affairs Minister Leon Schreiber told an Inter-Ministerial Committee on Migration briefing in Pretoria that the department expects a biometric online platform for asylum applications to be operational by 10 November 2026. According to SAnews, applicants will have to submit biometric information. The department will use it to verify identity, check for earlier rejected applications, authenticate passports and other travel documents, and flag other recorded infringements. The announcement is sensible administrative modernisation. It is also a large collection of sensitive data announced with no public description of how that data will be protected.
The strongest case for the platform
The argument for biometric intake is serious, and it should be stated fairly. An asylum system that cannot tell whether the person at the counter has already been refused will reward repeat filing. Schreiber said the platform is meant to stop rejected applicants from lodging fresh claims to delay removal. Duplicate and fraudulent claims consume scarce adjudicator time, and that time is taken from genuine refugees waiting in the queue. Biometric matching is the standard tool for this problem. Schreiber also pointed to the department's experience with the Electronic Travel Authorisation system as a model, and EWN reports he sees the same technology improving the asylum process.
The backlog gives the case urgency. Schreiber estimated the appeals backlog at about 120,000 cases, though he said he did not have the exact figure and that the department would confirm it. A different number appears in the parliamentary record. At a 26 May 2026 meeting of the Portfolio Committee on Home Affairs, the Refugee Appeals Authority reported a backlog of 161,000 cases, of which 70,976 were active and 90,024 inactive. The two figures come from different dates and sources, and the active/inactive split probably explains part of the gap. Even so, a public system that cannot state its own backlog with confidence has a data problem that this platform could help solve.
Where the argument gets weaker
The platform does not target the stage where delay actually builds up. SAnews reports that initial asylum applications take about 30 days to process and that most delays occur at the appeals stage. The department added 74 advocates for appeals, and Schreiber said the administration inherited a system where only nine people handled appeals. Biometric intake helps by weeding out duplicates at the front door. It does not clear 70,000 or more active appeals. If the platform is judged by the backlog, it will disappoint unless appeals capacity keeps growing alongside it.
The larger gap is data protection. South Africa's Protection of Personal Information Act (POPIA) treats biometric information as special personal information. Section 26 bars a responsible party from processing it unless an exception applies. Section 27 supplies those exceptions, including processing that is necessary for the establishment, exercise or defence of a right or obligation in law. Home Affairs will almost certainly have a lawful basis, and the Regulator can also authorise processing in the public interest, subject to conditions. The point is not that the platform is unlawful. The point is that a lawful basis is only the start of compliance.
What the public does not yet know is substantial. The EWN report describes no privacy protections or data-handling rules, and the other coverage reviewed here does not either. Open questions include how long biometric records will be retained, who inside government can query them, whether they will be shared with the enforcement systems being built elsewhere in the department, and what recourse an applicant has when a match is wrong. These are not minor details. Asylum seekers cannot realistically refuse to hand over data, so consent offers weak protection and published rules have to do that work instead.
The enforcement link
The asylum platform does not stand alone. In July, Home Affairs tendered for a biometric case management system for its inspectorate and 600 handheld biometric devices for field identity checks. Biometric Update reported bids closing on 24 July, a 36-month contract and a possible two-year extension. Its coverage focused on fraud detection and national security and did not mention privacy. Nothing in the sources reviewed says the asylum database and the enforcement tools will be connected. That is exactly why the department should say whether they will be. Biometric data collected to process a protection claim and then reused for field enforcement is a function-creep risk that a good design would rule out in writing.
What proportionate looks like
The fix is cheap and does not require delay. A proportionate approach would include four steps before or alongside launch:
- Publish a short processing notice setting out purpose, retention period and access controls for the asylum biometric records.
- State plainly whether asylum data will be shared with enforcement systems, and bar secondary use unless the law specifically allows it.
- Give the Information Regulator early sight of the design, so problems surface before millions of records exist rather than after.
- Create a human review route for false matches, since an erroneous flag could wrongly stall a genuine claim.
None of these measures weakens the platform's stated purpose of catching repeat filers and forged documents. They make it more defensible, including before the courts. The government also plans to seek a supervisory order from the Constitutional Court on its 7 July 2026 judgment, and a 29 October hearing in the Gauteng High Court will consider conflicting asylum judgments. A system that launches into that litigation with published safeguards is on firmer ground than one that does not.
The 10 November date is a target rather than a confirmed launch, and no reporting found after the announcement confirms it was met. Schreiber's central claim, that technology improves immigration processes, has support in the ETA experience. The remaining question is whether South Africa applies the same discipline to the privacy rules as it does to the software. Biometric systems tend to be easy to expand and hard to roll back, so the rules should be written first.