How US banks are using AI to analyze customer preferences, monitor workers, identify people sleeping near ATMs, and more, despite civil liberties concerns
Context & Ripple Effects
This Reuters report lands in a period when the big AI vendors were drawing lines around surveillance-adjacent work: by late 2021, the AI ethics chiefs at Google, IBM, and Microsoft said their companies had turned down client requests for credit scoring and facial recognition systems. Banks, which Evident later found staffing roughly 40% of some open roles with AI positions, kept building anyway — the Reuters piece maps where: customer preference analysis, worker monitoring, and even flagging people sleeping near ATMs.
The privacy tension is not new to financial services. Expensify's use of Mechanical Turk to process documents its automation couldn't handle set an early template for fintech privacy blowback, and the worker-surveillance thread here runs directly into today's emotion AI debate over facial and sentiment analysis tracking white-collar moods.
First-order effects
- Bank customers and employees are directly subject to AI profiling — preference scoring, worker monitoring, and ATM-area identification — with civil liberties concerns attached to each use case as reported by Reuters.
Second-order effects
- With major vendors like Google, IBM, and Microsoft rejecting facial recognition and credit-scoring requests, banks' demand shifts toward suppliers willing to build surveillance tools, and hiring pressure compounds it — JPMorgan alone advertised thousands of AI roles in early 2023.
Third-order effects
- If banks normalize mood- and behavior-tracking AI on workers and the public, financial services becomes a proving ground for workplace surveillance techniques now spreading into white-collar jobs, forcing regulators to define where customer analytics ends and civil-liberties violations begin.
The trend: Financial institutions are becoming the largest deployers of surveillance-grade AI on customers and workers, filling the gap left by big tech vendors' ethics-driven refusals.