“AI for the rest of us.” This phrase encapsulates Apple’s consumer-first approach to generative artificial intelligence (GenAI). Unlike the centralized AI strategies of other tech giants, Apple’s AI vision is rooted in user privacy and on-device processing. This presents a unique opportunity for banks to learn from Apple’s innovative approach and enhance their AI applications’ safety, security, and personalization.

In the first half of 2024, banks reported over $1.3 Billion in scam-related losses involving vishing (voice phishing), where scammers impersonate trusted individuals to extract personal information. Centralized systems often fail to detect such scams promptly, and banks currently need a reliable and scalable way of doing so. However, the model can analyze voice patterns and detect real-time anomalies with on-device AI, offering immediate and robust protection.

Another example is the recent outage in MS Windows due to inadequate patch releases. This highlights the vulnerabilities of centralized systems. In the future, as GenAI agents handle sensitive tasks like credit card transactions, banks need robust, safe, transparent, and decentralized AI solutions. Apple’s on-device AI approach is a shiny example of the future of AI in banking.

AI management remains complex, often filled with buzzwords and various interpretations. Banks face a strategic decision between focusing on core processing efficiency and delivering personalized customer experiences. A balanced approach, integrating cost-efficient processing with improved customer services, is not just essential, but a strategic necessity in today’s AI landscape.

Apple Intelligence, debuting in beta for iPhone 15 Pro and later devices, integrates advanced AI-driven features such as writing tools, an upgraded Photos app, and a more conversational Siri integrated with ChatGPT. Apple’s AI approach emphasizes privacy, on-device processing, and secure cloud computing, contrasting sharply with the data-driven methods of other tech firms.

Banks handle sensitive customer information and must ensure data protection at all costs. By adopting Apple’s on-device processing model, banks can minimize data transmission and storage risks, significantly reducing the potential for data breaches. On-device AI can process transactions, detect fraud, and provide personalized services without transmitting sensitive information to external servers. Today’s anomaly models are complex to maintain and manage in a decentralized fashion, but small LLMs installed on the device can detect anomaly patterns without increasing the computation requirement centrally. For instance, if a malicious actor takes over your account, the on-device AI can analyze typing speed and patterns to alert the bank of fraudulent activity.

Apple’s strategy of utilizing smaller, localized AI models can serve as a blueprint for banks aiming to enhance personalization in their AI systems. Large, centralized models often incorporate biases present in diverse training data. In contrast, smaller models installed on devices tailored to individual styles and preferences can provide personalized services such as travel itineraries, hotel bookings, and financial advice while preserving user privacy.

Transparency in AI operations is crucial for building customer trust. Apple’s practice of prompting users for consent before data sharing can be a model for banks. Implementing precise consent mechanisms ensures customers know how their data is used, fostering trust and compliance with regulations such as GDPR.

Apple’s approach to transparency involves clear and straightforward communication with users about how their data is used. Banks can adopt this practice by clearly informing customers how their data will be processed and stored, ensuring no hidden data practices. This openness can alleviate customer concerns and increase trust in the bank’s AI systems, especially as banks collaborate with external AI systems in real time.

Apple provides regular updates and reports on data management, a practice banks can emulate. Transparency reports should inform customers about AI decision-making processes, data usage, and third-party integrations. These reports should be accessible and comprehensible, helping customers stay informed and confident about their data security.

Transparency also involves regular ethical audits of AI systems to ensure fairness and accountability. Inspired by Apple’s transparency measures, banks should conduct routine audits of their AI models to detect and mitigate biases or unethical practices. Publishing audit results can enhance transparency and demonstrate the bank’s commitment to ethical AI usage.

Apple’s strategic integration of ChatGPT into Siri while maintaining user privacy through ‘private cloud computing’ demonstrates a balanced approach to leveraging external AI capabilities. Banks can adopt a similar strategy, integrating third-party AI services for advanced analytics and customer insights while ensuring data privacy and security are not compromised.

As AI regulations evolve, Apple’s privacy-centric approach provides a proactive framework for compliance. Banks adopting similar strategies will be better positioned to navigate regulatory challenges while maintaining customer trust.

Apple’s AI on the edge approach, characterized by on-device processing, smaller models, and a strong emphasis on privacy and transparency, offers valuable insights for the banking sector. Banks can improve their AI applications by prioritizing data security, enhancing personalization, ensuring transparency, providing safer, more secure, and more personalized customer services.

Following Apple’s lead can help banks build trust and credibility, driving wider adoption of AI technologies in the financial sector.

Our next blog will explore how AI on the edge could reduce and prevent fraud and scams.