Welcome to the inaugural edition of Neurians AI Digest, our curated newsletter on all things Artificial Intelligence (AI).

Our AI digests will cover insights & updates on how AI is impacting industries such as Healthcare, Education, Finance, Retail, Travel & the Software industry itself.

Here’s a recap of this week in AI:

Healthcare:

Microsoft on Thursday announced new health-care data and Artificial intelligence tools, including a collection of medical imaging models, a health-care agent service and an automated documentation solution for nurses. 

The tools aim to help health-care organizations build AI applications quicker and save clinicians time on administrative tasks such as documentation, a major cause of industry burnout.

“By integrating AI into health care, our goal is to reduce the strain on medical staff, foster the collective health team collaboration, enhance the overall efficiency of healthcare systems across the country,” Mary Varghese Presti, vice president of portfolio evolution and incubation at Microsoft Health and Life Sciences, said in a prerecorded briefing with reporters. (Source: CNBC)

Microsoft’s new AI capabilities aim to address the largely multi-modal nature of modern medicine that includes data sources such as medical imaging, genomics & clinical records.

The newly launched multi-modal capable Microsoft Healthcare AI models include:

• MedImageInsight: An embedding model that enables sophisticated image analysis, including classification and similarity search in medical imaging. These models enable healthcare organizations to automatically route imaging scans to specialists or flag potential abnormalities for further review.

• MedImageParse: This multi-modal model designed for precise image segmentation, can be fine-tuned for specific applications such as tumor segmentation or organ delineation enabling AI to assist with highly targeted cancer and other disease detection, diagnostics, and treatment planning.

• CXRReportGen: By incorporating current and prior images, along with key patient information, this multimodal AI model generates detailed, structured reports from chest x-rays, highlighting AI-generated findings directly on the images to align with human-in-the-loop workflows.

(Source: Microsoft announcement - Oct 10th, 2024)

Read more about Microsoft’s Healthcare AI announcements HERE.

A leading healthcare startup that builds artificial intelligence (AI) assistants in healthcare, Suki, which competes with Microsoft-owned Nuance announced that it has raised $70 million in a Series D round. (Source: Reuters)

Three scientists, [David Baker (Biochemist at University of Washington), Sir Demis Hassabis & John Jumper, Computer scientists at Google Deep mind] who discovered powerful techniques to decode and even design novel proteins - the building blocks of life- were awarded the Noble prize in Chemistry last Wednesday. Their work leveraged Artificial Intelligence (AI) & holds the potential to transform how new drugs and other materials are made.

The field of nutrition is being significantly disrupted by AI. Finding the perfect diet that works for anyone can be challenging. Enter AI to help with precision diet recommendations. (Source: Forbes)

Education: 

Visionary entrepreneur Elon Musk unveiled the “Optimus” robot at Tesla’s “We Robot” event on October 10th. The Optimus robot is revolutionary with a slated purchase price between $20,000 to $30,000. The Optimus robot could potentially serve as a “Teaching Assistant” learning and executing tasks such as preparing teaching materials and supervising students. This could significantly unburden K-12 teachers. (Source: Forbes)

While AI has tremendous potential to transform education, currently, there are no universal guidelines, leaving education companies to navigate issues such as data privacy, transparency, and the ethical use of AI on their own. In response, various ed-tech organizations and school districts are pushing for the development of responsible AI standards.

Several groups, including the EdSafe AI Alliance and the Consortium for School Networking, have created frameworks and rubrics that emphasize AI safety, fairness, and efficacy. These guidelines aim to help both vendors and schools assess the quality and ethical use of AI tools. However, the rapidly evolving nature of AI requires guidelines that are flexible and can adapt to new developments.

Federal regulations, such as the Kids Online Safety Act and COPPA 2.0, may also impact AI use in schools, especially regarding data privacy. School districts are increasingly incorporating specific AI-related requirements in their requests for proposals (RFPs), demanding transparency from vendors about how data is handled and used.

(Source: GovTech)

Finance

Despite it’s massive potential to transform the finance industry/function, only 9 % of respondents in a survey of various roles within the office of the CFO said that their organizations finance function was in the process of exploring ways to scale AI Projects more broadly. (Source: CFO Dive). Reasons for low adoption could include skill gaps, budget constraints or concerns about Data security.

However, A bullish outlook exists for the financial services industries projected spend on Artificial Intelligence (AI) with a projected rise from $35 billion in 2023 to $97 billion in 2027 at a CAGR of 27%. (Source: Forbes)

Four key areas where financial services (FS) firms are leveraging generative AI (gen AI) for immediate gains include:

1. AI Co-pilots: AI systems that assist employees in tasks like customer service, fraud detection etc. These co-pilots may eventually help in real-time investment strategies and market predictions.

2. Always-on AI web crawlers: These tools continuously monitor financial news, market movements, and social media sentiment, helping banks proactively manage risks and identify opportunities.

3. Automating unstructured data: Gen AI transforms unstructured data (e.g., emails and documents) into actionable insights, freeing employees for higher-value tasks like strategic decisions.

4. Hyper-personalization: AI helps create tailored services by analyzing financial and non-financial data. For instance, Klarna’s AI assistant has reduced marketing costs by 25% by proactively engaging with customers.

In the future, gen AI may enhance risk management through the use of synthetic data for better fraud detection and compliance. Fintechs are also driving democratization of AI for smaller institutions, enabling broader access to advanced AI tools. While Financial services firms have focused on efficiency, the next frontier will be using gen AI to drive growth and innovation.

(Source: Forbes)

Retail

With the massive holiday shopping season kicking off this month and black friday around the corner, major retailers like Walmart and Amazon are ramping up their AI capabilities to help enhance the shopper experience.

Features such as Amazon’s Rufus, a generative AI-powered shopping assistant designed to help customers make better-informed purchasing decisions by integrating information from Amazon's catalog and the web are gearing up to scale up for the peak holiday season this year. (Source: AWS)

To operate at scale and meet customer needs, Rufus relies on AWS infrastructure, including specialized AI chips (AWS Inferentia and Trainium), which provide high performance and low costs for managing large language models (LLMs).

Rufus's infrastructure needed to handle multi-billion-parameter models with low latency, especially during high-demand periods like Amazon Prime Day. The team deployed a system using AWS services such as Elastic Container Service (ECS) and Triton Inference Server, across multiple regions for scalability and resilience. This setup reduced latency and energy costs by 4.5 times compared to other solutions, while maintaining fast response times under 1 second.

Rufus uses Retrieval Augmented Generation (RAG), combining product data and customer queries to deliver accurate and reliable shopping assistance. To optimize performance, the team developed a real-time inference streaming architecture, reducing response time further. Additionally, innovations like continuous batching increased throughput and minimized hardware idling.

The system was optimized for scalability, achieving high hardware utilization and efficient load balancing through the Least Outstanding Requests (LOR) algorithm, significantly boosting throughput. The overall architecture allowed Rufus to handle over 80,000 AWS chips during Prime Day, supporting 3 million tokens per minute with low latency, while maintaining energy efficiency.

Another retail giant, Walmart is embracing "Adaptive Retail," a new era focused on creating personalized and convenient shopping experiences through cutting-edge technologies like AI, Generative AI (GenAI), Augmented Reality (AR), and Immersive Commerce. This strategy tailors shopping to individual customer preferences, both online and in physical stores. (Source: Walmart)

Key innovations include:

1. Retail-Specific AI Models (Wallaby): Walmart has developed proprietary GenAI platforms, including Wallaby, a series of large language models (LLMs) trained on Walmart's data. These models create personalized customer experiences, enhance customer service, and handle actions like managing returns.

2. Personalized Shopping: A new AI-powered Content Decision Platform predicts customer preferences to customize Walmart.com, creating unique homepages for each shopper. This personalization is expected to roll out in the U.S., Canada, and Mexico by next year.

3. Augmented and Immersive Commerce: Walmart's AR platform, Retina, creates 3D shopping experiences, with integrations into virtual environments like Roblox and ZEPETO. These immersive platforms allow users to shop for real and virtual items, enhancing engagement with younger generations.

Travel

Priceline has integrated AI-powered voice technology into its booking system, enhancing its Penny chatbot with OpenAI's GPT-4 API to enable real-time voice conversations. This update allows users to perform hands-free searches for hotels, get travel recommendations, and explore destinations via voice commands on Priceline's iOS app and website. The new feature aims to simplify the travel booking process and improve accessibility, especially for individuals with mobility or vision challenges. (Source: PYMNTS)

The travel industry, worth $766 billion globally, is increasingly adopting AI technologies to streamline operations and enhance customer experiences. Other companies like Expedia, Amadeus, Delta Airlines, and Hopper are also leveraging AI for trip planning, customer service, and price predictions. However, challenges remain in addressing complex travel needs such as multi-destination itineraries, group bookings, and special accommodations. Experts highlight potential issues with accuracy, particularly with names and accents, but Priceline is working on improving its system's understanding of diverse speech patterns and accents across 120 languages.

Despite these advances, human travel advisors are still considered essential for complex travel needs, and AI is expected to complement rather than replace human interaction in travel planning.

With the upcoming holiday travel season, improving airport and passenger experience is critical for the aviation industry. Nick Woods, CIO of MAG Group, which operates major UK airports is focused on making airports more intelligent by leveraging AI, data insights, and partnerships with tech companies. MAG is implementing AI to improve operational efficiency and enhance customer experiences. Woods’ approach involves working with startups and co-investing in emerging technologies to streamline processes, such as flight scheduling and passenger flow management. (Source: ZDNET)

MAG uses machine learning to predict flight schedules and optimize operations through platforms like Google Cloud and AWS. Real-time data from sensors and cameras help monitor passenger movements, improving internal processes like check-in and immigration management. AI also plays a key role in automating seasonal planning, traditionally done manually, by optimizing schedules and improving decision-making.

Woods aims to enhance customer services by integrating AI to provide passengers with real-time updates and personalized offers. MAG is working with AWS to explore AI applications like monitoring aircraft turnaround processes to ensure on-time performance. Overall, Woods emphasizes a process-driven approach, where technology is used to support well-understood business needs, shifting MAG towards a more collaborative and innovative future.

Software

The ultimate goal for AI companies is to achieve AGI-Artificial General Intelligence .

However, there appear to be different definitions of what this entails amongst the leaders in the AI technology platforms such as Anthropic and OpenAI.

Anthropic CEO, Dario Amodei prefers the term “Powerful AI” instead. He defined the goal as “a world driven by this emerging form of AI, one in which everything goes right”.

OpenAI’s stated goal is to “Develop AI systems that are generally smarter than humans”.

Amodei’s vision of powerful AI (which he believes will be possible by 2026) includes following properties (Source: Business insider):

• It's smart: It is smarter than a Nobel Prize winner across most relevant fields.

• It's multimodal: It isn't limited to one mode of action. It can interface through text, audio, video, mouse and keyboard control, and the internet.

• It's independent: It does not just passively answer questions, Instead, it can be given tasks that take hours, days, or weeks to complete, and then goes off and does those tasks autonomously.

• It's abstract: It does not have a physical embodiment.

• It's replicable: The resources used to train the model can be repurposed to run millions of instances of it.

• It's fast. The model can absorb information and generate actions at roughly 10x-100x human speed."

• It's cooperative. Each of these million copies can act independently on unrelated tasks, or if needed can all work together in the same way humans would collaborate.

Hope you enjoyed this weeks Neurians AI Digest! Till next week…