​AI Tech Circle​

Stay Ahead in AI with Weekly AI Roundup; read and listen on AITechCircle:

Welcome to the weekly AI Newsletter. This newsletter is a valuable resource for me and many others in the community. It provides practical and actionable ideas that can be immediately applied to our jobs and businesses.

Before we begin, please share this week’s updates with a friend or colleague:

Today at a Glance:

• Empowering and protecting children from AI
• Generative AI Use case: Drafting Contracts and Statements of Work (SoWs) in Procurement
• AI Weekly news and updates covering newly released LLMs
• Open Tech Talk Podcast, the latest episode, From Blockchain to Generative AI: Deep Dive with JB Benjamin

Generative AI: Opportunities & Risks for Kids

How does Generative AI affect our children? Today’s topic shifts from AI’s business applications to the potential impact of Generative AI on our young ones. As a parent of three, I recently researched this issue, spending six to eight hours. Here are the key findings from my investigation.

Some statistics on the usage of AI in kids:

• ​Unicef has reported that 58 percent of kids aged between 12-18 are using the ChatGPT
• another surprising outcome is that many hide the usage from the parents and teachers
• ​Ofcom reported 59% of 7-17 year-old and 79% of 13-17 year-old internet users in the UK.
• New research from Internet Matters shows that a quarter of children already use AI tools to assist with their schoolwork. Four in ten children have interacted with generative AI, with over half of 13-14-year-olds among them.

​

Generative Intelligence is reshaping the world and significantly influencing the lives of current and future generations of children. Today, children engage with AI through various technologies, such as toys, virtual assistants, video games, and adaptive learning software. AI algorithms suggest what videos children should watch, which news to read, what music to enjoy, and who to connect with.

Beyond these direct interactions, AI-driven automated systems also indirectly affect children’s well-being by influencing decisions related to welfare benefits, healthcare quality, education access, and housing applications for their families.

This impact is widespread, affecting children globally, including those in developing countries, who may face missed opportunities due to limited access to AI technology benefits.

The opportunities:

Generative AI opens new doors for children, offering tools beyond entertainment and education to reshape how they learn, create, and engage with the world.

The vast and impactful opportunities, from personalized learning experiences to enhanced healthcare support.

​

The Risks:

While Generative AI offers great potential for enhancing children’s learning and creativity, it also brings certain risks that must be carefully managed. Ensuring these technologies are used safely is critical to protecting children’s well-being and future opportunities.

​

Chatbots are the widespread and most common use case of Generative AI in different areas.

​

The human-like tone of chatbots blurs the line between animate and inanimate objects, raising concerns about their impact on children’s development and privacy.

The research shows these interactions may affect children’s understanding of intelligence, cognitive growth, and social behavior, particularly during crucial developmental stages.

​

Unicef’s Policy Guidance with 9 Requirements for Child-Centered AI:

​UNICEF created this guidance through an extensive consultation process involving input from various experts to address the local AI-related needs and realities of policymakers and businesses globally while also incorporating children’s perspectives.

Source: Policy guidance on AI for Children from Unicef

Another detailed “Artificial Intelligence for Children toolkit” is available from the World Economic Forum.

Weekly News & Updates…

Last week’s AI breakthroughs marked another leap forward in the tech revolution.

1. Yi-Coder 1.5B and 9B under Apache 2.0. It is a collection of open-source code large language models (LLMs) that achieve state-of-the-art coding capabilities with under 10 billion parameters. Offered in two sizes—1.5B and 9B parameters—Yi-Coder comes in both base and chat versions, optimized for efficient inference and adaptable training. Yi-Coder-9B, in particular, enhances Yi-9B by incorporating 2.4 trillion high-quality tokens, carefully curated from GitHub’s code repositories and code-related data filtered from CommonCrawl. Link​
2. Reflection Llama-3.1 70B is an open-source LLM trained with a new technique called reflection tuning. This technique teaches an LLM to detect mistakes in its reasoning and correct course. The model was trained on synthetic data generated by Glaive. Link​
3. Llama 3.1: 405B. These models deliver improved reasoning capabilities, a larger 128K token context window, and improved support for 8 languages, among other improvements. Link​

The Cloud: the backbone of the AI revolution

• Oracle Launches Secure Cloud Computing Architecture (SCCA) for Brokers and Integrators Serving the U.S. Department of Defense link​
• NVIDIA at Oracle CloudWorld 2024: Shaping the Future of AI and Data Processing link​

Gen AI Use Case of the Week:

Generative AI use cases in the Government and Public Sector :

Drafting Contracts and Statements of Work (SoWs) in Procurement, this use case is derived from Deloitte.

This use case involves using large language models (LLMs) to streamline your procurement process by automating contract and SoW drafting with AI. Discover how LLMs can save time, reduce legal risks, and drive efficiency in your organization.

Business Challenges

1. Legal teams and procurement departments spend significant time and resources drafting contracts and SoWs, especially for complex projects.
2. Ensuring compliance, minimizing errors, and aligning with local regulations can be tedious and time-consuming.
3. The volume of contracts for larger organizations makes scaling manual review and drafting challenging

AI Solution Description

Using Large Language Models (LLMs), organizations can automate the initial drafting of contracts and SoWs by providing critical inputs such as project scope, objectives, and terms. LLMs trained in legal terminology and procurement-related content will generate customized drafts that align with company policies and industry standards. Legal professionals can then review and fine-tune these drafts, significantly reducing the time and effort involved in the process. The LLMs would also enable clause extraction and template adaptation based on similar previous contracts, improving accuracy and efficiency.

Expected Impact/Business Outcome

• Revenue: By speeding up contract generation and reducing legal bottlenecks, companies can finalize deals and projects more quickly, increasing revenue potential.
• User Experience: Legal and procurement teams will be able to focus on higher-value activities, improving job satisfaction and output quality.
• Operations: Streamlined processes in contract generation will lead to faster approvals and reduced waiting times for departments reliant on procurement contracts.
• Process: AI-automated contract generation will ensure contract consistency and help reduce human errors, mitigating legal risks.
• Cost: By reducing manual legal work, companies can save on labor costs and allocate resources more efficiently.

Required Data Sources

• Historical contracts, SoWs, and procurement documents
• Industry-standard legal templates
• Internal company policies and guidelines
• Local and international legal regulations and standards for compliance
• ERP applications

Strategic Fit and Impact

Automating contract drafting with LLMs will significantly reduce administrative and legal overhead, speed up processes, and enhance compliance, making it a strategic investment for larger organizations. It aligns with a broader push toward digital transformation and operational efficiency in procurement and legal departments.

Rating: High Impact & strategic fit

Favorite Tip Of The Week:

Here’s my favorite resource of the week.

• DiffusionKit: Run Diffusion Models on Apple Silicon with Core ML and MLX. link​

Potential of AI

• GuideLLM is an advanced tool designed to assess and optimize large language model (LLM) deployments. It simulates real-world inference workloads, allowing users to evaluate performance, resource requirements, and cost considerations across different hardware setups. This ensures that LLMs are deployed efficiently, scalable, and cost-effectively without compromising service quality. link​

Things to Know…

OpenAI co-founder Sutskever’s new safety-focused AI startup SSI raises $1 billion, and it’s just an announcement on the HTML website. Link​

The Opportunity…

Podcast:

• This week’s Open Tech Talks episode 142 is “From Blockchain to Generative AI: Deep Dive with JB Benjamin,” JB is not just an entrepreneur but an exceptionally technically proficient lead UI/UX developer. His imaginative approach has led to the development of multiple successful digital businesses.

Apple | Spotify | Youtube

Courses to attend:

• ​AI Python for beginners from Deep Learning: Learn Python programming fundamentals and how to integrate AI tools for data manipulation, analysis, and visualization.

Events:

• ​Oracle Cloud World, Sep 9-12, Las Vegas, USA
• ​GITEX GLOBAL, Oct 14-18, 2024, Dubai, UAE
• ​EUROPEAN Conference on Artificial Intelligence, Oct 19-24, 2024 Santiago de Compostela
• ​TED Conference on AI, October 17-19, 2024 | Vienna, Austria

Tech and Tools…

• ​AnythingLLM: Chat with your docs and use AI Agents, which are hyper-configurable and multi-user.
• ​Perplexica is an open-source AI-powered searching tool or an AI-powered search engine that goes deep into the internet to find answers

Data Sets…

• ​PAWS-X: This dataset contains 23,659 human-translated PAWS evaluation pairs and 296,406 machine-translated training pairs in six typologically distinct languages: French, Spanish, German, Chinese, Japanese, and Korean.
• ​ShapeNet: ShapeNetCore is a subset of the full ShapeNet dataset with single clean 3D models and manually verified category and alignment annotations. It covers 55 common object categories with about 51,300 unique 3D models.

Other Technology News

Want to stay on the cutting edge?

Here’s what else is happening in Information Technology you should know about:

• Does A.I. Really Encourage Cheating in Schools? as reported by The New Yorker​
• Ant Group launches AI-powered ‘life assistant’ app reported by finextra​

Join a mini email course on Generative AI …

​Introduction to Generative AI for Newbies​

Earlier week’s Post:

That’s it!

As always, thanks for reading.

Hit reply and let me know what you found most helpful this week – I’d love to hear from you!

Until next week,

Kashif Manzoor