Guidelines from the European Union's Anti-Money Laundering Authority (AMLA) on automated profiling are drawing sharp criticism from data-protection advocates, Heise reported on 16 October, with critics warning that algorithmic risk-scoring of bank customers could amount to building a surveillance state.
AMLA, based in Frankfurt, is assembling the EU's single rulebook and supervisory framework for anti-money-laundering compliance — a job that increasingly means deciding how far obliged institutions may lean on automated systems. Those systems score customers, transactions and relationships for the probability of illicit activity, which is the only practical way for banks to process the data volumes involved.
The objection is not that automation is inherently illegitimate. It is that the safeguards around it are thin: an algorithm may flag a person as high-risk without a clear explanation, the flagged status can be hard to contest, and the practical burden of being wrongly suspected falls on people with the least power to challenge it.
For advocates, normalising opaque scoring in the name of financial-crime detection sets a template that other domains — benefits, immigration, policing — can copy. For supervisors, the alternative is banks drowning in transactions they cannot meaningfully review by hand. How AMLA lands that trade-off will shape privacy rules well beyond finance.
