Gen AI for Business #3

Welcome to this week’s edition of Gen AI for Business! As always, we’ve curated a selection of the most compelling stories and insights from across the generative AI landscape to help you stay ahead in this rapidly evolving field. This issue packs diverse topics, from new tools that accelerate AI application development to thought-provoking discussions on the ethical and practical challenges of integrating AI in various industries to inspire your strategies.

If you enjoy reading this newsletter, please comment and share!?

Knowledge is power.

Eugina

B2B News of the Week: models, industries, and the battle for hosting

News about models and everything related to them

The Cohere Toolkit expedites generative AI app development for scalable cloud deployments, while FrugalGPT enhances LLM cost-effectiveness. Snowflake launches Snowflake Arctic, a versatile LLM, and Apple announces an AI-driven Safari browser. SK Telecom is developing an advanced LLM tailored for telecoms, aiming to surpass GPT-4's performance with specialized training. MIT researchers have developed three neurosymbolic methods that enhance large language model performance in coding, AI planning, and robotics by using natural language, significantly improving task accuracy.?

• 6 Latest AI Models from Meta, OpenAI, Apple and More Recent advancements in open AI models from major tech companies include Meta's Llama 3, Snowflake’s Artic, Microsoft’s Phi-3 Mini, Megalodon, Mistral AI’s Mixtral 8x22B, and Apple's development of its own generative AI, dubbed "Apple GPT.

• The Clear And Present Danger of Open LLMs ?concerns about the risks open LLMs may pose, such as misinformation, privacy issues, and the manipulation of these models for harmful purposes. The piece calls attention to the need for careful consideration and possibly regulation as these powerful AI tools become more integrated into various sectors of society and everyday applications.

• New Cohere Toolkit Accelerates Generative AI Application Development ? The new Cohere Toolkit is designed to expedite the development of generative AI applications across any cloud provider so developers can deploy in production at scale.?

• ?? Implementing FrugalGPT: Reducing LLM Costs & Improving Performance discusses "FrugalGPT," a framework designed to optimize the usage of large language models (LLMs) for cost-effectiveness and improved performance. It outlines strategies like prompt adaptation to reduce processing costs, LLM approximation using model fine-tuning, and LLM cascade to choose the most cost-effective model based on input. These methods aim to achieve similar or better performance than high-cost models like GPT-4 but at a significantly reduced expense.

• Snowflake has announced the launch of Snowflake Arctic, an LLM that the company describes as “the most open, enterprise-grade LLM on the market.” Snowflake's Open-Source Arctic LLM Supports Diverse Enterprise Requirements - Acceleration Economy ?

• Google Gemini AI -- all you need to know about it. ?

• Apple to unveil AI-enabled Safari browser alongside new operating systems Apple plans to release an AI-enhanced Safari browser and new operating systems, emphasizing user privacy and enhanced AI capabilities for tasks directly from the browser, like making phone calls from web pages. This is part of Apple's strategy to integrate powerful on-device AI while maintaining privacy by minimizing off-device data processing.?

• This news is personally exciting to me. I have been in telecom for 24 years, even started a new market category and have 12 patents. The industry moves slow, so this is exciting: SK Telecom to unveil LLM for telecom carriers in June . Eric Davis, vice president of AI tech collaboration at SK Telecom, also previously mentioned that the new LLM would perform about 35 percent better compared to the GPT-4. SK Telecom has repeated the process of training the model with communication-related data, supervising fine-tuning between general models and custom models, evaluating whether it provides useful answers to inquiries and learning insufficient data to enhance the efficiency and professionalism of its LLM, according to SK Telecom officials.

• Natural language boosts LLM performance in coding, planning, and robotics | MIT News MIT researchers at CSAIL have developed three neurosymbolic methods that enhance the performance of large language models (LLMs) in coding, AI planning, and robotics by utilizing natural language for better abstraction representation and to perform more like humans. The frameworks, named LILO, Ada, and LGA, each address different aspects: LILO for code synthesis and documentation, Ada for sequential decision-making in AI agents, and LGA for improving robotic interaction with environments.

• Microsoft, Apple look to go big with smaller AI models | Semafor shows the shift in the AI industry toward developing smaller AI models, exemplified by Microsoft's Phi-3 and Apple's OpenELM. These models, designed for broader device compatibility like smartphones and sensors, use less computing power compared to larger models like OpenAI's GPT-4. Microsoft aims to optimize performance in smaller models, while Apple's OpenELM, though more limited, integrates with its iPhone ecosystem. The development of these models emphasizes quality data use over quantity.

• LLM deployment flaws that catch IT by surprise – Computerworld ? When deploying large language models (LLMs), companies are finding some big security loopholes. These AI systems might skip important safeguards or access too much data without showing signs of a crash, which can lead to unintended breaches or errors. It's a reminder for all in IT to rigorously monitor and test these tools, ensuring they're not just powerful but also safe and effective.

• Unpacking the Potential Risks of Generative AI Chatbots on Local Government Websites - Coates’ Canons NC Local Government Law ? Local governments are exploring the use of generative AI chatbots for public communication on their websites. While these chatbots offer advanced interactive capabilities by generating responses from large language models, they pose significant risks due to their potential to produce inaccurate or misleading information. This unpredictability could lead to legal liabilities, especially if incorrect information influences public actions. It's crucial for local governments to consider these risks carefully and implement robust verification systems if they choose to deploy such AI tools.

Industry New Partnerships

Apple renews talks with OpenAI for iPhone generative AI features, Bloomberg News reports | Reuters ?

Gen AI news from different industries

President Biden's executive order enhances AI safety and equity in healthcare, demanding data transparency and non-discrimination. Natural language processing aids in MDRO screening; generative AI in patient portals boosts efficiency but needs human oversight to ensure accuracy and manage biases. The NIST standardizes generative AI assessments for safety and ethics across sectors. Banking discussions emphasize critical thinking over AI reliance, while a roadmap for small businesses promotes secure, ethical AI integration. The military focuses on trustworthy AI for strategic uses, real estate leverages AI for analytics and customer relations, and the legal sector balances AI efficiency with ethical considerations.

Healthcare?

• What the Biden administration’s artificial intelligence executive order means for radiology and it references this: President Biden issues ‘landmark’ executive order on artificial intelligence ? The Biden administration's AI executive order aims to ensure responsible AI use in healthcare and radiology. It introduces stricter safety and equity standards for AI tools, mandates the disclosure of AI training data, and requires adherence to non-discrimination laws. The order also anticipates more significant federal oversight and regulation of AI applications in healthcare.?

• Natural language processing can inform real-time MDRO screening Natural language processing can analyze patient history and physical notes. A study found that it may be able to help hospitals make MDRO screening decisions.

• Gen AI patient portal messages OK, but need human review ? Generative AI (Gen AI) is proving useful in drafting responses for patient portal messages, helping to reduce the workload for healthcare providers. However, these AI-generated responses still require a human review before being sent to ensure accuracy and appropriateness. The technology, while efficient, carries risks such as perpetuating biases, thus necessitating oversight by medical professionals to maintain the quality and empathy in patient communication.?

• Harnessing AI for clinical data registries: A guided approach Gen AI implementation for clinical data registries needs robust governance and though there is potential of generative AI to enhance operational and clinical efficiency, lack of formal policies governing AI use is a challenge.We need to establish clear objectives and continuous evaluation of AI technologies to ensure they meet healthcare goals while safeguarding patient privacy and data.

?Government

NIST, the U.S. agency known for setting technical standards, has introduced a new platform designed to evaluate generative AI technologies. This platform aims to provide detailed assessments of generative AI applications, ensuring they meet the required standards of safety, reliability, and ethics. This move is part of a broader effort to establish benchmarks for emerging technologies and to foster a standardized approach to their development and deployment. GenAI - Evaluating Generative AI ?

Banking

The discussion highlights a trend where reliance on AI, such as using ChatGPT for instant solutions, may undermine the value of deep expertise and critical thinking in professional settings. This critique points out that while AI levels the playing field by making advanced tools accessible, the real competitive edge lies in the ability to apply critical thought to the outputs AI generates. So, yes, we still need humans! Bank CIO: We don't need AI whizzes, we need critical thinkers to challenge AI | ZDNET ?

Small Business

This article provides a roadmap for small businesses to safely and securely integrate AI into their operations. It emphasizes the importance of understanding AI capabilities and potential risks, prioritizing data security, and ensuring AI ethics and transparency. The guidance suggests that businesses should start small, with clear goals for AI usage, and gradually scale up their AI initiatives while continuously monitoring for security and ethical compliance. This approach helps small businesses harness AI's benefits while managing associated risks effectively. A Small Businesses Roadmap For Safe And Secure AI ?

Military

Military halts use of generative AI The Pentagon is actively working to enhance trust in its military AI systems by ensuring their responsible use. This includes a strategic focus on making AI dependable and ethical within defense operations. The goal is to integrate AI in a way that supports superior decision-making on the battlefield while maintaining high standards for reliability and accountability.?

Real Estate

How Real Estate Professional Can Utilize Generative AI Key applications include using AI for predictive analytics to forecast market trends, optimize pricing strategies, and manage properties more efficiently. AI also supports automated valuation models for quicker property appraisals. Additionally, real estate agents leverage AI-driven tools for customer relationship management, enhancing client engagement through personalized communication and recommendations based on data-driven insights.

Legal

Checking the Pulse: Bird's Eye View on Gen AI - Showcases, Reviews, and Reactions in the Legal World The article discusses the integration of generative AI in legal practices, highlighting both the advantages, such as enhanced efficiency and cost reduction, and challenges, including ethical concerns and the potential for biases in AI outputs. It advises legal professionals to start small with AI integration and stay informed about new developments and regulations to effectively and ethically leverage AI technologies.

Commercial aviation

How Generative AI Will Change the Commercial Aviation Industry ? Generative AI will significantly influence the commercial aviation industry by improving flight planning, maintenance, and operational efficiency. It will integrate into key areas like revenue management and predictive maintenance, enhancing productivity and customer service while addressing challenges such as regulatory compliance and data privacy.

Shipping and Logistics

UPS is leveraging generative AI to enhance customer interactions and improve logistics efficiency. By automating customer message responses, UPS has reduced the workload on their agents and paved the way for broader applications of AI technologies. This initiative is part of their broader strategy to optimize package delivery and improve the customer experience by analyzing vast amounts of data and utilizing sophisticated AI models. UPS delivers customer wins with generative AI | CIO ?

Gen AI Market Trends

Generative AI funding reached $25.2 billion in 2023, with leaders like Apple and Nvidia shaping the AI PC sector, and Intel advancing enterprise AI. This growth is fueled by innovations like diffusion models, expanding applications into audio synthesis and computer vision, and integrating AI across various sectors.

• Generative AI Funding Hits $25.2 Billion in 2023, Report Reveals

• Apple, Nvidia Lead AI PC Sectors, Enterprise Market up for Grabs

• And Intel is trying to grab it: Intel welcomes Open Platform for enterprise AI - Elettronica Plus ?

• The Future of Generative AI: Trends, Challenges, & Breakthroughs ? This growth is underpinned by innovations like diffusion models, making generative AI more energy and cost-efficient. Applications are expanding beyond simple text generation to include areas such as audio synthesis, computer vision, and more. With technology becoming mainstream, its integration into business, education, and entertainment continues to evolve, presenting both new opportunities and challenges.

?

What/where/how Gen AI solutions are being implemented today?

Chatbots are key AI tools in business, while advanced tools like Sora still require human intervention for complex tasks. Deloitte notes a readiness gap in fully integrating AI in enterprises, emphasizing the need for human oversight. Measuring ROI for generative AI is complex, impacted by its broad effects on business beyond simple cost metrics. Nvidia plays a role in enhancing enterprise generative AI applications, focusing on personalizing customer interactions, automating coding through AI tools like StarCoder2, and utilizing collaborations with Snowflake to develop tailored AI models, all aimed at improving efficiency and reliability in business operations.

• https://www.forbes.com/sites/quickerbettertech/2024/04/26/chatbots-emerge-as-the-first-killer-ai-app-for-businesses/?sh=34eb69ca77b9 ?

• Creators of Sora-powered short explain AI-generated video’s strengths and limitations | TechCrunch ? While Sora allows for the creation of realistic and dynamic videos, the process is not fully automated and still requires significant traditional filmmaking techniques like storyboarding, editing, and post-production work, including color correction and visual effects. The filmmakers found that while the tool is impressive, achieving precise control over aspects like character actions and scene consistency is challenging and often requires creative workarounds. Sora's outputs sometimes include unwanted elements, necessitating additional edits. This highlights that while AI tools like Sora are advancing, they are not yet at a stage where they can independently produce polished films without human intervention.

• Organizational readiness is not there yet:

From Delloite State of Generative AI in the Enterprise 2024 | Deloitte US Q2 report, data from February 2024.??

• Five Myths About Generative AI That Leaders Should Know - Knowledge at Wharton These myths include the ideas that generative AI can fully automate complex job roles without human oversight, that it can seamlessly integrate into existing business processes without significant adjustment, and that it inherently understands ethical or cultural contexts. The piece emphasizes the importance of recognizing the limitations of AI and the need for strategic implementation and human oversight to maximize its benefits effectively.

• How to measure ROI: Why Measuring the ROI of Transformative Technology Like GenAI Is So Hard ? The primary difficulties stem from defining clear metrics for success, quantifying the impact of AI on various aspects of business operations, and the unpredictability of long-term outcomes. For GenAI, the direct financial returns may not be immediately apparent, making it harder to establish a straightforward ROI calculation. Additionally, the impact of GenAI often extends beyond simple cost savings or revenue generation, influencing factors such as innovation, competitive advantage, and customer experience, which are harder to quantify but can be crucial for long-term success.

• Nvidia: The Most Valuable Enterprise Gen AI Use Cases | Technology Magazine 1. The use of AI to create personalized customer experiences, where businesses can leverage AI-powered tools to tailor marketing and services based on detailed customer data analysis.2. the development of AI-driven coding tools, such as StarCoder2, a collaborative effort by Nvidia, ServiceNow, and Hugging Face, which automates coding and enhances developer productivity. 3. The partnership between Nvidia and Snowflake focuses on using Snowflake's data cloud for developing custom AI models, for Data management and AI modeling, improving business applications like chatbots.3. composite AI, which integrates various AI technologies to solve complex problems more effectively, emphasizing the importance of robust, reliable AI systems for enterprise use.?

Key Investments and Financial Highlights for Gen AI

• The Week’s 10 Biggest Funding Rounds: Xaira And Other AI Startups Have Huge Week Recent funding rounds in AI and biotech have shown significant investor interest, with Xaira Therapeutics securing a $1 billion investment for AI-driven drug discovery and Augment receiving $227 million for AI coding assistance. These investments reflect confidence in AI's potential to revolutionize both biopharmaceuticals and software industries, suggesting a future where AI integration can drive major advancements and efficiency in business operations.

• Google Invests Billions in Data Centers, AI Skills Development ? New $3 billion data center will be built in Indiana and $75 million set aside to fund AI skills training.

• Amazon’s generative AI business has now hit a multibillion-dollar run rate that’s reaccelerated cloud growth | Fortune This milestone reflects the increasing integration of AI technologies across various sectors, contributing heavily to AWS’s revenue.??

• Meta plans to build $800 million, next-generation data center in Montgomery ? Meta is building an $800 million data center in Montgomery, Alabama. This facility, designed for AI optimization, will span 715,000 square feet and is expected to create 100 jobs. It will aim for LEED Gold certification and run on 100% renewable energy. Construction will start soon with completion targeted for late 2026.

The battle for cloud business?

• Google Cloud revenue soars as AI drives billions in infrastructure spend | CIO Dive ?

• Oracle boosts its generative AI capabilities as cloud competition heats up ?

Regional updates on Gen AI: bills, regulations, patents, and partnerships

SenseTime's SenseNova 5.0 outperforms OpenAI's GPT-4, showcasing China's AI leadership. The U.S. advances AI safety and equity under President Biden's executive order, introducing new standards and a security board. Meanwhile, Microsoft's $1.7 billion investment in Indonesia aims to enhance AI capabilities and digital infrastructure and also $2.2B to bring AI and cloud technologies to Malaysia. Japanese Prime Minister Fumio Kishida introduced an international framework to regulate generative AI.?

• Japanese Prime Minister Fumio Kishida introduced an international framework to regulate generative AI. Launched at the OECD in Paris, this initiative, named the Hiroshima AI Process, has been joined by 49 countries. It aims to create a safe and trustworthy AI environment globally by addressing risks such as disinformation and promoting international cooperation on AI standards. Japan’s Kishida unveils a framework for global regulation of generative AI | The Seattle Times ?

• Chinese startup SenseTime announced SenseNova 5.0, an AI model that beats OpenAI’s leading GPT-4 Turbo model across almost every key performance benchmark. China's SenseTime Unveils Sense Nova 5.0, Surpassing OpenAI's GPT-4 in AI Benchmark Tests

• China's AI industry is rapidly advancing, as showcased at the 2024 Zhongguancun Forum in Beijing. A highlight is "Tong Tong," a virtual AI avatar designed for proactive assistance in various industries such as emergency services and healthcare. China has become a major player in AI, with a core industry worth approximately 70.37 billion USD and over 4,400 related enterprises. The nation is also a leader in large language models, with a significant number of self-developed models and innovative applications across different sectors. China Focus: China embraces AI boom, diverse application scenarios-Xinhua ?

• Role of the US government The U.S. Needs to ‘Get It Right’ on Artificial Intelligence | TIME ? It stressed the importance of the U.S. leading in safe, secure, and equitable AI development while considering both the immense potential and the ethical challenges of AI technology. The discussion underlined that getting AI policy right is crucial for maximizing AI's benefits across society while managing its risks, particularly in areas like national security, privacy, and workforce impacts. The emphasis was on ensuring that AI advances do not outpace the country's ability to manage them responsibly.?

• Biden-Harris Administration Announces Key AI Actions 180 Days Following President Biden’s Landmark Executive Order | The White House ? This includes developing a screening framework to prevent AI from creating harmful biological materials and establishing an AI Safety and Security Board. Additionally, draft documents have been released for public feedback on managing generative AI risks, with efforts to set international standards. The administration also addresses AI's potential for discrimination in various sectors and boosts AI employment in federal agencies.

• The department’s National Institute of Standards and Technology (NIST) has released four draft publications intended to help improve the safety, security and trustworthiness of artificial intelligence (AI) systems. GenAI - Evaluating Generative AI ? NIST has also launched a challenge series that will support development of methods to distinguish between content produced by humans and content produced by AI. Department of Commerce Announces New Actions to Implement President Biden’s Executive Order on AI | NIST ?

• Responsible Use of Generative Artificial Intelligence for the Federal Workforce (US)? The Office of Personnel Management (OPM) has released guidance on using Generative Artificial Intelligence (GenAI) in federal work. While GenAI can improve tasks like drafting, coding, and translation, it carries risks like copyright infringement and bias. Agencies must authorize its use and implement human checks to review outputs for accuracy and biases before dissemination. This guidance follows the Biden administration's executive order on AI, ensuring responsible and equitable AI deployment across federal agencies.?

• Number of AI Patents granted annually by region shows that China is the winner:

• Who in Europe is investing the most in artificial intelligence? | Euronews —This article might explain the lack of EU patents in the EU, it highlights the significant commitment from the European Union to enhance AI capabilities through the Next Generation EU fund, which allocates €4.4 billion to AI initiatives. Italy and Spain lead in direct investment amounts, with Italy planning to spend €1.895 billion on AI-related projects. The ongoing challenges remain regulatory hurdles and skill shortages.

• Microsoft will invest $1.7 billion in AI and cloud infrastructure in Indonesia | AP News ? Microsoft is expanding its investment in Indonesia, focusing on AI and digital transformation. This includes establishing the first datacenter region in Indonesia to enhance local cloud services, data security, and privacy. The initiative, part of the "Berdayakan Ekonomi Digital Indonesia" project, also aims to significantly upscale digital skills among Indonesians. This strategy is intended to boost Indonesia's digital economy and facilitate widespread access to AI technology across various sectors.

• Microsoft announces a US$2.2 billion investment to fuel Malaysia’s cloud and AI transformation. This investment aims to advance new cloud infrastructure and AI technologies within the country, part of a broader initiative to transform Malaysia into a digitally advanced economy. This aligns with Malaysia's goals to improve technological accessibility and skills among its population, fostering innovation and economic growth through high-tech upgrades and AI integration across various sectors.

Learning Center

Workforce readiness (Deloitte report).?

Below are resources that can help you and your teams geat educated on how you can use Gen AI in your business.

• Recording from LInkeidn Live on health and wellness: https://www.youtube.com/watch?v=LpyqVTaUyeA&t=8s ?

• NVDIA video on adoption of Gen AI in healtcare. It highlights how accelerated computing and AI technologies are transforming medical and life sciences through faster diagnoses, advanced medical instruments, and new treatment methodologies. Key innovations include AI-driven insights for disease treatment, surgical assistance, and drug discovery, emphasizing AI's role in enhancing healthcare delivery and patient care.?

• Register for? the 3rd session on Gen AI in B2B -- we are discussing implementation in the Supply Chain https://www.dhirubhai.net/events/7191108732812996608/comments/ ?

• Register for the 4th session on Gen AI in B2B – how sw engineering organizations can implement Gen AI. https://www.dhirubhai.net/events/7191475534604771328/comments/ ?

• SLM, LLM and hybrid AI models: Bigger isn’t always better: How hybrid AI pattern enables smaller language models - IBM Blog ?

Prompt of the week

ChatGPT can help you with the right prompt for your project. Ask “Can you suggest a prompt for me to do a particular task?”

Tools and Apps

• If you wondered, how AI apps make money … most still use traditional subscription models, with some incorporating usage-based elements. Trends include a common use of free versions to boost initial adoption and varied degrees of pricing transparency. There is room for innovation in AI app monetization, particularly through outcome-based pricing models that could enhance customer adoption and revenue. How AI apps make money - by Kyle Poyar and Palle Broe ?

• For authors, drafts, brand voice, notes: https://strut.so ?

• Generative AI Powered Assistant - Amazon Q - AWS Amazon Q is a generative AI-powered assistant from AWS, designed to integrate with enterprise systems to automate tasks, generate content, and provide data insights. It offers features like secure connectivity to business tools, customizable responses, and privacy-focused controls, ensuring compliance with corporate data governance. Amazon Q supports diverse business functions through its ability to analyze data, generate reports, and assist in decision-making processes, making it a versatile tool for enhancing productivity across various organizational roles.

Source: AWS

• Customer interactions and internal ops: Recent updates in generative AI tools from companies like AWS, RingCentral, Qualtrics, and Avaamo focus on enhancing customer interactions and internal operations. AWS's Amazon Connect now summarizes customer conversations post-contact. RingCentral's new RingEX platform includes AI-driven note-taking and summary features. Qualtrics integrates dynamic AI agents for automated response handling, while Avaamo improves its conversational AI for more context-aware interactions. No Jitter Roll: Another Slew of Gen AI-related News from AWS, RingCentral, Qualtrics and Avaamo ?

• Free AI Diagram Generator - EdrawMax AI ?

• ?Truva.ai Truva helps software platforms onboard and retain customers with AI agents.?

• Altern – search for new AI tools with https://altern.ai/ ??

• Harvey – platform for legal professionals.

• Introducing the Claude Team plan and iOS app \ Anthropic ( Claude uses "constitutional AI" to ensure ethical responses and reduce harmful outputs. Claude supports various tasks like language processing and analysis, and is available via API and chat interfaces for both developers and businesses. The models—Haiku, Sonnet, and Opus—vary in performance to suit different needs, with Haiku being the fastest and Opus the most complex. Additionally, Claude prioritizes privacy by not retaining user data, making it ideal for sensitive environments).

• Open Source LLM engineering platform.

• Build a Custom LLM with ChatRTX | NVIDIA (running on your PC)?

• AI-driven contract analysis and management Agiloft launches Gen AI Prompt Lab including ‘bring your own LLM’ capability - Legal IT Insider ?

• GPT-4, ChatGPT and AI Detector tool by ZeroGPT ?

• And for fun, as you plan your summer travel: Mindtrip ??

?If you enjoyed this newsletter, please comment and share. If you would like to discuss a partnership or to invite me to speak at your company or event on Gen AI for business, please DM me.

P.S. I tried creating a “Newsletter Editor” GPT to help me with the newsletter, and it did not go as planned. So, it’s me, the human, curating it. I will experiment again next week and will report back. But the avatar is cool. Don’t you think so?