Mandatory Quarterly Compliance Disclosures by Market Intermediaries for AI/ML Tools: SEBI

Jan 20, 2019 | Maathangi Hariharan

Mandatory quarterly reporting for all AI/ML system based applications by market intermediaries

Starting 2019 on an optimistic note, the Securities and Exchange Board of India (SEBI) issued a recent circular directing all market intermediaries offering AI/ML system based applications and services to mandatorily supply compliance reports every quarter to SEBI. Effective March 31, 2019, all market intermediaries offering services that rely on AI/ML systems are required to participate in the reporting system and submit quarterly reports within 15 days from the completion of a quarter.

Covering wide ground, this circular addresses trading and investing, offering investment strategy, advice and compliance activities that deploy AI/ML systems. According to Annexure B of the circular, SEBI has identified an indicative list of AI/ML systems, whose use is subject to mandatory reporting by market intermediaries. This includes the use of Natural Language Processing (NLP), Neural Networks and systems that maintain a knowledge base. Under Annexure A, the concerned market intermediary is also required to inform SEBI among others, a description and manner of use of the AI/ML system and the type of safeguards in place to prevent any abnormalities that may arise.

To address the possible incidences of misrepresentation arising from advertising financial benefits, SEBI has identified AI/ML systems as ‘black boxes’ whose behaviour cannot be easily quantified. As ‘black boxes,’ there is limited understanding of the internal workings of an AI system. By identifying AI/ML systems put to use by market intermediaries to offer financial services, SEBI is adopting an innovation friendly, yet precautionary stance in protecting consumer interest. It recognizes the limitations that ‘blackboxes’ currently pose (Srivats Shankar offers in-depth analysis on the same, which can be found here), in terms of the need for accountability and opacity i.e. inability to understand how an AI/ML system makes decisions and why it makes such decisions.

As the Indian market regulator, SEBI has adopted a very forward looking approach to innovation centric financial services that are gaining traction.

Related

Latest

European Union Considering Establishing International Treaty on Artificial Intelligence

Srivats Shankar | Jul 26, 2020

The European Union Ad Hoc Committee on Artificial Intelligence is considering an international treaty for artificial intelligence in its upcoming session

Department of Transportation Releases Document for Future of Transportation Regulatory Approach

Srivats Shankar | Jul 26, 2020

The US Department of Transportation releases document for facilitating communication between regulators and developers of emerging transportation technology

Amica: Australia’s Solution for Dividing Assets

Srivats Shankar | Jul 23, 2020

Read On

Terminology

Deepfake

/diːpfeɪk/

Artificial General Intelligence

/ˌɑːtɪfɪʃl ˈdʒɛn(ə)r(ə)l ɪnˈtelɪɡəns/

Artificial Intelligence

/ˌɑːtɪfɪʃl ɪnˈtelɪɡəns/

More Terminology

Events

13th International Conference on Agents and Artificial Intelligence – ICAART 2021

INSTICC | 2021-02-04 12:00:00

More Events