August 04, 2024
Kannan Subbiah
vCxO | IT Governance | Mutual Funds | Independent Director | C|CISO | CGEIT | CISA | FCA
It’s both logical and tempting to design your AI usage around one large model. You might think you can simply take a giant large language model (LLM) from your Act 1 initiatives and just get moving. However, the better approach is to assemble and integrate a mixture of several models. Just as a human’s frontal cortex handles logic and reasoning while the limbic system deals with fast, spontaneous responses, a good AI system brings together multiple models in a heterogeneous architecture. No two LLMs are alike — and no single model can “do it all.” What’s more, there are cost considerations. The most accurate model might be more expensive and slower. For instance, a faster model might produce a concise answer in one second — something ideal for a chatbot. ... Even in its early days, gen AI quickly presented scenarios and demonstrations that underscore the critical importance of standards and practices that emphasize ethics and responsible use. Gen AI should take a people-centric approach that prioritizes education and integrity by detecting and preventing harmful or inappropriate content — in both user input and model output. For example, invisible watermarks can help reduce the spread of disinformation.
One of the superpowers LLMs bring to the table is ultra-efficient summarization. Given a dense information block, generative AI models can extract the main points and actionable insights. Like with our earlier trials in algorithmic root cause analysis, we gathered all the data we could surrounding an observed issue, converted it into text-based prompts, and fed it to an LLM along with guidance on how it should summarize and prioritize the data. Then, the LLM was able to leverage its broad training and newfound context to summarize the issues and hypothesize about root causes. Constricting the scope of the prompt by providing the LLM the information and context it needs — and nothing more — we were able to prevent hallucinations and extract valuable insights from the model. ... Another potential application of LLMs is automatically generating post-mortem reports after incidents. Documenting issues and resolutions is not only a best practice but also sometimes a compliance requirement. Rather than scheduling multiple meetings with different SREs, Developers, and DevOps to collect information, could LLMs extract the necessary information from the Senser platform and generate reports automatically?
So-called "AI toothbrushes" have become more common since debuting in 2017. Numerous brands now market AI capabilities for toothbrushes with three-figure price tags. But there's limited scientific evidence that AI algorithms help oral health, and companies are becoming more interested in using tech-laden toothbrushes to source user data. ... Tech-enabled toothbrushes bring privacy concerns to a product that has historically had zero privacy implications. But with AI toothbrushes, users are suddenly subject to a company's privacy policy around data and are also potentially contributing to a corporation's marketing, R&D, and/or sales tactics. Privacy policies from toothbrush brands Colgate-Palmolive, Oral-B, Oclean, and Philips all say the companies' apps may gather personal data, which may be used for advertising and could be shared with third parties, including ad tech companies and others that may also use the data for advertising. These companies' policies say users can opt out of sharing data with third parties or targeted advertising, but it's likely that many users overlook the importance of reading privacy policies for a toothbrush.
When it comes to technological alignment between banks and tech partners, it’s about more than ensuring tech stacks are compatible. Cultural alignment on work styles, development cycles and more go into making things work. Both partners should be up front about their expectations. For example, banking institutions have more regulatory and administrative hurdles to jump through than technology companies. While veteran fintech companies will be aware and prepared to move in a more conservative way, early-stage technology companies may be quicker to move and work in more unconventional ways. Prioritization of projects on both ends should always be noted in order to set realistic expectations. For example, tech firms typically have a large pipeline of onboarding ahead. And the financial institution typically has limited tech resources to allocate towards project management. ... Finally, when tech firms and financial institutions work together, a strong dose of reality helps. View upfront costs as a foundation for future returns. Community banking and credit union leaders should focus on the potential benefits and value generation expected three to five years after the project begins.
Specifically designed to be the Army’s forward biometrics collection and matching system, NXGBCC has been designed to support access control, identify persons of interest, and to provide biometric identities to detainee and intelligence systems. NXGBCC collects, matches, and stores biometric identities and is comprised of three components: a mobile collection kit, static collection kit, and a local trusted source. ... The Army said “NXGBCC will add to the number of biometric modalities collected, provide matches to the warfighter in less than three minutes, increase the data sharing capability, and reduce weight, power, and cost.” NXGBCC will use a Local Trusted Source that is composed of a distributed database that’s capable of being used worldwide, data management software, forward biometric matching software, and an analysis portal. Also, NXGBCC collection kit(s) will be composed of one or more collection devices, a credential/badge device, and document scanning device. The NXGBCC system employs an integrated system of commercial-off-the-shelf hardware and software that is intended to ensure the end-to-end data flow that’s required to support different technical landscapes during multiple types of operational missions.
AI models are becoming increasingly powerful while also getting smaller and more efficient. This advancement enables them to run on edge devices without compromising performance. For instance, Qualcomm’s latest chips are designed to handle large language models and other AI tasks efficiently. These chips are not only powerful but also energy-efficient, making them ideal for mobile devices. One notable example is the Galaxy S24 Ultra, which is equipped with Qualcomm’s Snapdragon 8 Gen 3 chip. This device can perform various AI tasks locally, from live translation of phone calls to AI-assisted photography. Features like live translation and chat assistance, which include tone adjustment, spell check, and translation, run directly on the device, showcasing the potential of edge computing. ... The AI community is also contributing to this trend by developing open-source models that are smaller yet powerful. Innovations like the Mixture of Agents, which allows multiple small AI agents to collaborate on tasks, and Route LLM, which orchestrates which model should handle specific tasks, are making AI more efficient and accessible.