IOSCO proposes six measures to assist members in creating appropriate regulatory frameworks to supervise market intermediaries and asset managers that use AI and ML.
IOSCO (International Organization of Securities Commissions) has proposed new guidance to help its members regulate and supervise the use of AI (artificial intelligence) and ML (machine learning) by market intermediaries and asset managers.
While the use of AI and ML may benefit firms and investors – such as by increasing execution speed and reducing the cost of investment services – it may also “create or amplify risks, potentially undermining financial markets efficiency and causing harm to consumers and other market participants,” IOSCO says.
The IOSCO Board identified AI and ML as an important priority last year to protect against potential risks and prevent consumer harm. The guidance was included as part of IOSCO’s annual work programme for 2020.
From engagement with market intermediaries and asset managers, IOSCO identified six key areas of potential risks and harms in relation to AI and ML: governance and oversight; algorithm development, testing and ongoing monitoring; data quality and bias; transparency and explainability; outsourcing; and ethics.
In the new consultation report, IOSCO proposes six measures to assist members in creating appropriate regulatory frameworks to supervise market intermediaries and asset managers that use AI and ML, namely around:
- appropriate governance, controls, oversight and accountability frameworks over the development, testing, use and performance monitoring of AI and ML
- robust, consistent and clearly defined development, testing, and monitoring processes to enable firms to identify potential issues prior to full deployment of AI and ML
- ensuring staff have adequate knowledge, skills and experience to implement, oversee, challenge the algorithms and outcomes of AI and ML, and conduct due diligence on any third-party providers
- third party providers relationship management, including to monitor and oversee performance and ensure accountability over outsourced functions
- appropriate transparency and disclosures to investors, regulators and other relevant stakeholders
- controls in place to ensure that the data that AI and ML performance is dependent on is of sufficient quality to prevent bias and sufficiently broad
Although the guidance is not binding, IOSCO members are encouraged to consider the proposed measures carefully in the context of their legal and
The report also includes a chapter on guidance for AI and ML published by supranational organisations such as the IMF (International Monetary Fund) and FSB (Financial Stability Board).
The consultation report, available here, is open for comment until 26 October 2020.