🚗 Yummy drive-thru

PLUS: The first proper AI legislation is officially here

Haroon ChouderyAug 2, 2024 · 8 min read

New here? Welcome to NAB, and make sure to subscribe to stay informed on the latest developments in AI!

Happy Friday, fellow humans 👋

The AI drive-thru trend is back this year, Google announced three new additions to their Gemma 2 AI model family, and the European Union's landmark AI Act has officially come into force…

Let’s dive into it!

Have anything you’d like to share with over 40k AI enthusiasts?

Sponsor Not A Bot

🧵 In today's edition:

🚗 Yummy drive-thru
🤖 Bigger isn’t always better
📜 The first proper AI legislation
🛠️ AI Tools to Check Out
🤑 AI Fundraising News
🐦 Tweet Post of the Day

🚗 Yummy drive-thru

The AI drive-thru trend is officially back this year, with Yum Brands, the parent company of Taco Bell, announcing plans to implement AI-powered drive-thru ordering at hundreds of Taco Bell locations across the United States by the end of 2024.

Here’s the TL;DR:

The company has already successfully tested the technology in over 100 Taco Bell restaurants, reporting improvements in order accuracy, reduced wait times, and decreased employee workload. Yum Brands aims to expand this technology globally in the future.
While other restaurant chains like Wendy's and White Castle have also been exploring AI drive-thru systems, Taco Bell's rollout is notably more extensive.
The fast-food industry is cautiously adopting this technology, balancing the potential benefits of improved efficiency and cost savings against customer and employee satisfaction.

Yum Brands' Chief Innovation Officer, Lawrence Kim, expressed confidence in the technology's effectiveness after two years of testing and refinement. The company is also testing voice AI at five KFC locations in Australia.

Overall: This development comes as part of a broader trend in the fast-food industry to leverage AI technology for enhancing operations and customer experience. Yum Brands is set to report its second-quarter earnings soon, which may help us understand the impact of this technology on its business.

Read more: CNBC

🤖 Bigger isn’t always better

This week, Google announced three new additions to their Gemma 2 AI model family, focusing on responsible AI development:

Gemma 2 2B: A new 2 billion parameter model that offers exceptional performance for its size, outperforming larger models on the LMSYS Chatbot Arena leaderboard. It's designed for flexible deployment across various hardware, from edge devices to cloud systems.
ShieldGemma: A suite of safety content classifier models built on Gemma 2. These classifiers target four key areas of harm: hate speech, harassment, sexually explicit content, and dangerous content. They're available in different sizes to suit various needs and performance requirements.
Gemma Scope: A model interpretability tool using sparse autoencoders (SAEs) to provide unprecedented insight into Gemma 2's decision-making processes. It includes over 400 freely available SAEs, interactive demos, and an easy-to-use repository.

Overall: These additions aim to support researchers and developers in creating safer AI experiences, gaining a deeper understanding of model behavior, and deploying powerful AI responsibly.

The new tools and models are openly available under commercially friendly terms, reflecting Google's commitment to transparency and collaboration in AI development.

By prioritizing safety, accessibility, and transparency, Google is working towards a future where AI benefits everyone while addressing key concerns in responsible AI development.

Read more: Google Blog

📜 The first proper AI legislation

Source: Consultancy.eu

The European Union's landmark AI Act has officially come into force, marking the world's first comprehensive regulation aimed at governing AI development and use, applying a risk-based approach to different AI applications.

Here are the big takeaways:

Strict obligations for "high-risk" AI applications, such as autonomous vehicles and medical devices.
Bans on "unacceptable-risk" AI uses like social scoring systems and workplace emotional recognition tech.
Specific requirements for general-purpose AI, including generative AI models.
Potential fines of up to 7% of global annual revenues for breaches.

The law will significantly impact U.S. tech giants like Microsoft, Google, and Meta, who are heavily invested in AI development. It applies to any organization operating or impacting the EU, regardless of location.

While the law is now in effect, most provisions won't be enforced until at least 2026. GenAI systems have a 36-month transition period to achieve compliance.

Overall: The EU AI Act sets a global precedent for AI regulation, similar to how GDPR influenced data privacy practices worldwide.

It aims to balance innovation with ethical concerns and potential risks associated with AI technology.

Read more: CNBC

🛠️ AI Tools to Check Out

Datrics: Your own AI Data Analyst that understands your data and delivers accurate, verified answers - designed to empower both technical and business users with the data insights they need. Check it out!
Klee: Local and secure AI on your desktop, ensuring comprehensive insights with complete data security and privacy. Check it out!
Tern: AI that creates a personalized trip itinerary in seconds. Check it out!

🤑 AI Fundraising News

When raises $4.6M in Seed funding to provide an AI-powered post-employment platform built for workers in between jobs who need support with health insurance, 401(k) rollovers, financial assistance, and career guidance.
Bee raises $7M in Seed funding to create a wearable device that uses AI to achieve the autonomy and memory required to be a reliable assistant.