Tech Insights AI 0005, Thursday 22nd August 2024

📰AI News Rundown

1.⚖️California's SB 1047: Navigating the AI Regulation Landscape

2.💸Anthropic's Breakthrough: Prompt Caching for Claude Cuts Costs and Latency

3.🇮🇳🚀India Accelerates AI Mission: Government Invites Bids for AI Infrastructure Providers

4.📱Protecting Children's Data in AI Apps: A Compliance Guide👶

5.📜AI Regulations Tracker Africa 🌍

6.💰AI Investments

7.🛠️ AI Tools

1.⚖️California's SB 1047: Navigating the AI Regulation Landscape

California's Senate Bill 1047, known as the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act, is stirring significant debate across the tech industry. The bill, introduced by State Senator Scott Wiener, aims to establish safety protocols for the development and use of large AI models, seeking to prevent potential "critical harms" such as the misuse of AI for creating weapons or orchestrating large-scale cyberattacks.

🔑Key Provisions of SB 1047

SB 1047 mandates that developers of large AI models, defined as those costing over $100 million and using 10^26 FLOPS during training, implement safety measures, including an "emergency stop" capability. The bill also requires annual third-party audits to ensure compliance with safety standards. Additionally, it proposes the creation of the Frontier Model Division within the Department of Technology to oversee these AI models and enforce civil penalties for non-compliance.

💬Industry Reactions

The bill has polarized opinions among tech companies, startups, and venture capitalists:

Article adapted from click here and here

2.💸Anthropic's Breakthrough: Prompt Caching for Claude Cuts Costs and Latency

Anthropic has introduced prompt caching for its Claude AI models, a feature that drastically enhances efficiency by reducing costs by up to 90% and latency by up to 85% for lengthy prompts. This innovation, now in public beta for Claude 3.5 Sonnet and Claude 3 Haiku, allows developers to cache frequently used context between API calls, optimizing applications like conversational agents and coding assistants.

⏱️Efficiency and Real-World Impact

Early adopters, such as Notion, have integrated prompt caching to make their AI features faster and more cost-effective, significantly improving user experience. The pricing model is designed to be economical, with cached prompts costing only 10% of the base input token price.Anthropic's prompt caching is poised to become an essential tool for developers seeking to streamline AI interactions and reduce operational costs.

3.🇮🇳🚀India Accelerates AI Mission: Government Invites Bids for AI Infrastructure Providers

The Indian Ministry of Electronics and Information Technology (MeitY) has issued a Request for Empanelment (RFE) to select AI service providers as part of the INR 10,372 crore India AI Mission.

The initiative aims to build a robust AI ecosystem in India by providing high-speed AI computing infrastructure to academia, startups, researchers, and government bodies at competitive rates. The mission focuses on creating a scalable AI infrastructure with over 10,000 GPUs, fostering innovation, enhancing datasets, and developing AI skills.

The empaneled service providers will offer cloud-based AI services for 36 months, with the possibility of extension. The initiative is part of a broader effort to harness AI's potential while addressing concerns about its misuse. Bid submissions are due by September 6, 2024

Understanding the Tender: What You Need to Know

4.📱Protecting Children's Data in AI Apps: A Compliance Guide👶

In the age of AI, safeguarding children's privacy is more important than ever. Legal Nodes ,latest article provides essential guidance for developers navigating the complex landscape of privacy laws.

Regulations Covered From understanding GDPR and COPPA requirements to designing child-friendly apps, this resource is a must-read for anyone involved in creating AI applications for young users. Prioritizing children's best interests and ensuring robust data protection measures are key to staying compliant and protecting vulnerable users.

🚨AI applications that have violated child protection laws⚖️

1. TikTok: TikTok was fined €345 million by Irish regulators for violating children's privacy under the EU's General Data Protection Regulation (GDPR). The violations were related to how TikTok handled children's data, including issues with age verification and privacy settings. Accounts for users aged 13 to 17 were made public by default, which was a significant concern.

2. Google Play Apps: Research by Comparitech found that about 20% of the top 500 kids' apps on Google Play were likely violating the Children's Online Privacy Protection Act (COPPA). These apps were collecting personal data from children without proper privacy policies or parental consent, which is required under COPPA

 📜Essential Resources for Ensuring Compliance in Child Data Processing✅

5.📜AI Regulations Tracker Africa 🌍

6.💰AI Investments

7.🛠️ AI Tools

Attention Tech Get Cross-Call Insights in Seconds You don’t have to study a dozen calls and emails to understand what your customers or prospects are thinking or how your team is performing.Attention does that for you.

WebAI establishing the first distributed Al infrastructure dedicated to personalized Al.

Exactly AI collaborate with the best illustrators discover models for any design needs and have it your way

Ori.co Serverless Kubernetes For AI Build Faster Scale Better Ori Serverless Kubernetes combines powerful scalability, simple management and affordability to help you train, deploy and scale your AI/ML models.

Hamming.AI Launch trustworthyAI voice apps in weeksPrompt optimization, automated voice testing, monitoring and more.Test your AI voice agent against 1000s of simulated users in minutes.

^{Thank you so much !}

It means a lot that you’ve read my newsletter and I hope it’s been informative for you and your team.

Would love to say hi? Here are the best places to find me

Share your feedback via email 📩 [email protected]