By Piyush Agarwal, SE Leader, Cloudera
The rise of new creative AI algorithms like large language models (LLMs) from OpenAI’s ChatGPT, Google’s Bard, Meta’s LLaMa, and Bloomberg’s BloombergGPT have led to increased awareness, interest, and adoption of AI across various industries. According to a 2023 report by PwC India, 54% of companies have implemented AI for business 1 . However, in highly regulated industries, the focus is less on off-the-shelf generative AI and more on understanding the relationship between their data and how AI can transform their business.
AI in Financial Services
A report by PWC India states that 57% of Indian financial institutions strongly agree that AI will give them a competitive edge over their peers 2 . AI has enabled the automation or augmentation of complex decision-making processes, personalized client experiences, the creation of customised customer education materials, and the matching of appropriate financial and investment products to individual customers’ needs. It represents a revolutionary technological development with significant potential.
However, this advancement is not without risks. Financial institutions must design AI systems that are transparent, reliable, fair, accountable, and compliant with privacy and security requirements.
Need for ‘Trusted AI’
Generative AI systems, like ChatGPT and others, pose a challenge due to the complexity of the datasets
used and their tendency to generate non-factual information. Additionally, there’s a risk of data leakage,
including intellectual property (IP) and personally identifiable information (PII) when using public SaaS-
based models, emphasising the importance of robust data encryption standards and clear ownership
clauses.
Generative AI and LLMs, unlike traditional narrow AI, have the ability to produce original, creative outputs by learning from input data, using a special form of neural network architecture called a
transformer model. However, integrating these AI solutions requires reshaping core business processes,
transforming application workflows, and adapting corporate cultures, presenting significant challenges
for financial institutions.
To mitigate risks associated with commercial AI services, the concept of Trusted AI is introduced. Trusted AI emphasizes integrating AI solutions using open-source technologies on a hybrid data platform that ensures control, flexibility, and protection of proprietary assets and sensitive data. By utilising AI models trained on the financial institution’s secure data and deploying them internally or in secure cloud environments, organisations can achieve transparency and maintain compliance with security and privacy standards unique to the financial sector.
AI Maturity Model to secure financial institution
For an effortless integration of AI, financial and insurance companies should invest in the AI Maturity model outlined below for a comprehensive and robust security strategy which includes the stages of
adoption, from foundational AI integration to fully mature AI integration.
1. Foundational AI Integration
At this foundational stage, financial institutions begin by prioritizing open-source AI tools and deploying a hybrid data platform. According to the Enterprise Data Strategy MarketPulse Survey for Cloudera by Foundry, 92% of the respondents agree that a hybrid cloud data strategy is important in today’s environment and gives organizations a competitive advantage, whereas 61% of them have already deployed a hybrid data platform. Along with deployment, businesses should start with basic process automation and chatbots by tapping on open-source LLMs while initiating AI training programs for employees.
2. Intermediate AI Integration
At this stage of adoption, financial institutions and insurance companies should leverage a foundational hybrid data platform to fully harness AI’s potential. Their focus should be on improving user experience, enabling data-driven decisions, and fortifying cybersecurity by enhancing customer interaction, automating loan and credit decision-making, leveraging AI for financial crime prevention, systematising governance, establishing feedback mechanisms, and facilitating communication between stakeholders.
3. Advanced AI Integration
In this phase, financial institutions and insurance companies deeply immerse themselves in AI, extracting valuable insights from data with the aid of the hybrid platform’s automation capabilities. Advancements include predictive analytics, regulatory compliance simplification, improved risk management, extensive AI training, future scalability planning, and automated bias testing to ensure ethical AI models.
4. Transformative AI Integration
With a strong open-source base and a functional hybrid data platform, AI is deeply integrated into an institution’s core processes. Stringent security measures guarantee authorized access to AI models and data. Monitor market trends, enhance cybersecurity, personalize customer experiences, automate processes, and facilitate stakeholder communication for successful AI integration.
5. Fully Mature AI Integration
At full maturity, financial institutions and insurance companies realize the power of Trusted AI built on top of a hybrid data platform, accelerating AI operationalization, with Trusted AI embedded across all operations. Businesses should implement advanced AI oversight, develop next-gen personalized financial products, utilize AI for real-time risk management, automate wealth management, anticipate regulatory changes, and explore cross-industry integration.
To unlock valuable insights from their data and gain a competitive edge, financial institutions should
partner with industry experts and solution providers to identify the capabilities needed to effectively navigate the complexities of the highly regulated environment in which they operate. This ability to handle large-scale data infrastructures securely and reliably, wherever they reside will be fundamental to ensuring the smooth adoption of AI and Generative AI technologies.