HomeEverythingEducationTV
Equities & FundsCrypto & Digital AssetsAI & TechnologyBusiness & CorporateUS Politics & PolicyGeopolitics & Global RiskMacro, Rates & FXCommodities & EnergyEuropean Politics & MarketsAsia-PacificReal Estate & Property
Story archiveAll categories
← All Stories

A third of banks do not maintain logs for GenAI models

Created at 14 Jul · 3:36 AM1 source↑ Market-relevant
IN SHORT

A Risk Benchmarking study found that only one specialist European lender registers and reviews complete production logs for its generative AI models, indicating a gap in model risk management practices across the banking sector.

✉Newsletter

PiQ Daily

Pick your topics. Get only what matters, on your cadence.

Key Numbers

one-thirdbanks not maintaining GenAI logs

Who's Involved

Risk Benchmarking
conducted a study on bank model risk management
specialist European lender
the only firm reviewing complete GenAI logs

↳ Why This Matters

The lack of systematic logging for generative AI models in a third of banks raises concerns about model risk management, regulatory compliance, and the potential for undetected biases or errors in AI-driven financial operations.

Key facts

  • A study revealed that one-third of banks do not systematically log production prompts and outputs for generative AI models.
  • Only a single specialist European lender registers and reviews complete records for its GenAI models.
  • Larger financial institutions are prioritizing higher-risk applications of GenAI.
  • The findings are part of Risk Benchmarking's inaugural Model Risk Management (MRM) study.

A recent study by Risk Benchmarking has highlighted a significant gap in the management of generative artificial intelligence (GenAI) models within the banking sector, with a third of surveyed institutions not maintaining systematic logs of production prompts and outputs. The inaugural Model Risk Management (MRM) study found that only one specialist European lender registers and reviews complete records for its current GenAI models.

Larger firms participating in the study indicated a focus on higher-risk use-cases for GenAI, suggesting a strategic approach to deployment despite the logging deficiencies. The findings underscore a broader challenge in ensuring robust oversight and accountability for AI technologies as they become more integrated into financial services.

Frequently asked questions

The study found that a third of banks do not systematically maintain logs for their generative AI models, with only one specialist European lender reviewing complete production prompts and outputs.

Logging is crucial for model risk management, allowing for review of prompts and outputs to ensure compliance, identify biases, and understand model behavior.

Yes, larger firms are reportedly focusing on higher-risk use-cases for GenAI, suggesting a more targeted deployment strategy.

What Happens Next

01Banks are expected to address logging deficiencies for GenAI models.
02Further studies will likely track improvements in AI model risk management practices.

Get the newsletter.

Pick the topics you actually care about. We'll email when there's news worth your time, on the cadence you choose. Cancel any time from your account.

Cadence

How It Developed

A study found that a third of banks do not maintain logs for generative AI models.
Only one specialist European lender registers and reviews complete production prompts and outputs for its GenAI models.
Larger firms are focusing on higher-risk use-cases for GenAI.

Sources

T1
A third of banks do not maintain logs for GenAI modelsRisk.net

Related Stories

China Develops AI Safety Benchmark Amidst Regulatory Focus on Large Models
13 Jul · 1:36 PM
Canada regulator warned banks on AI cyber risks, email shows
13 Jul · 6:07 PM
Wall Street banks ramp up digital assistants in bid to win productivity race
13 Jul · 9:06 AM
AI homework tools cut exam scores by 20%, study finds
14 Jul · 1:05 AM
Claude's Personality Varies by Model Version and Language, Anthropic Finds
13 Jul · 8:46 PM