☕️ 🟣 Data-Driven Enterprise • Green AI • Global Digital Governance • Algorithmic Art
Welcome to DecodingAI®, a newsletter on the business and governance of AI in financial services. This is a newsletter for Board Directors, C-suite executives, their teams and advisers.
1.Business
1.1 Strategy
This McKinsey guide is a useful companion for executives to help them understand the characteristics of the new data-driven enterprise and what enterprise capabilities they enable. It also details how to embed these characteristics in the enterprise. Link
1.2 Leadership
This article provides a moment for reflection on what moral values are and why we need to restore them in the leadership. I believe restoring moral values is imperative, especially for organisations undergoing profound digital transformation projects. This is also the topic of a research we are
conducting. This article poses deep questions frequently asked by boards, employees, business partners and even technology vendors:
In whom do we now recognise and salute exemplar leadership qualities?
Who are representative of great leadership?
What accounts for the growing diminution of leadership standards? Link
1.3 Fintech
This is a use case of applied AI in audit, as deployed by Cherry Bekaert, a top 25 accounting firm in the United States that has helped thousands of clients achieve their financial, operational and strategic goals. Link
1.4 Digital Assets
I keep returning to the topic of Central Bank Digital Currency (CBDC) which routinely forms part of the panel discussions and Q&A sessions I have been invited to speak on. I wrote a short piece with a few immediate thoughts and hard questions that I submitted for discussion on a panel of last week. Link
1.5 Military Technology
I found this China Aerospace Studies Institute report an informative read. It addresses military robotics, unmanned aerial systems and new capabilities for “new era” of Chinese military power.
1.6 Geopolitics
The year 2025 will be a landmark for digital diplomacy and global governance. It is the year of wrapping up the UN cybersecurity OEWG and the negotiations on cybercrime at the Ad Hoc group. 2025 will also flag the ‘last mile’ toward the realisation of the Agenda 2030 and 17 SDGs that are increasingly dependent on digital developments. This article provides a checklist that aims to help diplomats navigate global digital governance. Link
2. Governance
2.1 Cybersecurity
This Interpol cybersecurity strategy paper makes for easy and useful reading with many transferable points to the private sector in financial services. Link
2.2 Policy
The CEO of the UK’s financial services regulator (the FCA) recently gave a talk when he identified critical issues that the FCA is targeting. These issues have been identified as the cost of living crisis; the new consumer protection; enforcing the sanctions against Russia; economic, social and governance (ESG) reforms; diversity and inclusion; accountability and performance; innovation and crypto; post-Brexit regulation and the FCA’s future strategy. This is useful as it fleshes out the regulatory emphasis and provides guidance to fintech, crypto firms and the financial services incumbents where to focus their attention in the next 12 months, and beyond. Link
2.3 Ethics
Digital group thinking and prediction algorithms undermine our ability to think independently and critically. This is because digital group thinking creates a “reality” that your mind is mostly unable to challenge as untrue. An example is the power of social media in the wrong hands. This is also because we are relying on algorithms to decide the optimal route to achieve our objectives while using very little of own critical reasoning to understand or even challenge the validity of algorithmic recommendations. AI alters the human mind is a statement to reflect on. Link
2.4 ESG Data
Banque de France recently launched this brief note which is eloquently written. It contains useful diagrams that might help focus the discussion on finding solutions to the persistent ESG data gap. Financial stakeholders report ESG data gap as the key challenge to managing climate-related risks. Link
3. Technology
3.1 AI Research: Green AI
Transformer-based architectures are the model of choice for natural language understanding, and yet they come at a significant cost because they are complex (quadratic complexity). In the pursuit of Green AI, this paper investigates simple MLP-based architectures and identifies them as a route to efficiencies in respect of processing time, training data, and hyperparameter tuning. Link
3.2 Universal Dependencies
The Universal Dependencies (frequently referred to as UD) is a framework for morpho-syntactic annotation of human language, which to date has been used to create treebanks for more than 100 languages and has been widely adopted in 1950. It has since grown to become an open source international cooperative project, despite its many weaknesses and challenges, one being identified as a lack of focus on the meaning rather than the statistical value of grammar. In the 1980s, UD was identified as an unsolvable and remains so. Despite this, Google and Stanford remain staunch supporters which can only undermine the adoption of alternative solutions. The author of this article submits that UD are fundamentally flawed and provides a few useful ideas and new approaches to UD. Link
4. Learning
This book came recommended. It addresses deep learning in a visual way. I am yet to finish reading it, but I have read enough of it to recommend it to those who are interested in expanding their deep learning knowledge. It is an easy way to navigate and learn a complex topic. Link
Stats of the Week
AI-driven forecasting to supply chain management can reduce errors by between 20 and 50 % and reduces lost sales and product unavailability of up to 65%. Source
Diagram of the Week
This is useful diagram to refer to when trying to identify the next steps in employee upskilling strategy.
Art
We have reached this stage much sooner than many predicted: artists are protesting against AI generated art in a campaign titled "Stop AI Stealing the Show". Link
These are some examples of AI generative art or frequently referred to algorithmic art.
About me
My name is Clara Durodié. My expertise is at the intersection of technology, governance, and business in financial services. I work with leading financial and technology boards as well as their risk management and strategy teams. I am one of the first practitioners to pioneer Ethical AI in financial services.
If you liked this newsletter, then you will likely enjoy reading my book Decoding AI in Financial Services which you can buy here and read the reviews here.
If you would like me to present to your board, C-suite, investment committee or speak at your conference on the topics discussed above, please enquire with Blair Nichols bbn@bbncreativemanagement.com. More about my work here.
Did you learn something new from this newsletter? If so, please recommend it to your friends. Share the knowledge and help us grow.
If you would like to sponsor this newsletter, please email me: ea@cognitivefinance.ai
Decoding AI® Newsletter Disclaimer