The GIANT AI Report

The fast-paced AI landscape may seem overwhelming, as we are constantly inundated with new developments and breaking news. The GIANT AI Report is here to break it down into digestible bite-sized pieces that are currently relevant to parenting and education.

In this edition, we take a look at how AI is advancing towards emulating how children learn language, teaching and testing Next Generation Science Standards, synthesizing math theorems and proofs, and decoding ancient sea scrolls! Open AI partners with a university and establishes a new Child Safety Team. Semafor launches an AI-curated news feed. 


AI Learns Language Through the Experience of a Single Child in Groundbreaking Study

In a groundbreaking study published in the Science journal, researchers unveiled a machine learning model that emulates how children learn language, offering profound insights into early language acquisition. Using video and audio recordings from a child's perspective, the model successfully associated words with visual objects, shedding light on the intricate process of language understanding and use. The study challenges traditional theories and suggests that simple associative learning mechanisms, coupled with multimodal representation learning, can serve as a foundation for replicating the initial stages of word learning. 

Read more


Game Platforms Like Roblox Can Be a Breakthrough Teaching Tool

Educators and developers are exploring the integration of gaming elements into school assessments to make them more engaging, stealthy, and informative, echoing the immersive experiences of video games. Initiatives like the Northwest Evaluation Association's 3D-digital assessment on Roblox aim to leverage gameplay to measure complex concepts like Newton's Second Law of Motion. Despite challenges in mindset and assessment design, advocates envision a future where playful assessments not only capture a broader array of skills but also foster creative and inquiry-based learning environments. This shift signifies a departure from traditional test formats towards more interactive and empowering evaluation methods.

Read more 


AlphaGeometry: An Olympiad-level AI system for geometry

Math has been a known limitation for AI. However, in a recent report published in Nature a study presents AlphaGeometry, a novel theorem prover for Euclidean plane geometry, which addresses the challenge of automating mathematical theorem proving, especially in geometry where translation difficulties inhibit machine-learning approaches. Unlike previous methods relying on human demonstrations, AlphaGeometry synthesizes millions of theorems and proofs across varying complexities, leveraging a neuro-symbolic system guided by a neural language model trained on extensive synthetic data. Impressively, AlphaGeometry outperforms previous methods on olympiad-level problems, producing human-readable proofs and demonstrating the potential for automated reasoning to approach human-level performance in mathematical domains.

Read more


Elon Musk Backs AI Project to Decode Ancient Roman Scrolls

Elon Musk's backing of the Vesuvius Challenge, a project aimed at decoding ancient Roman scrolls damaged by the Pompeii eruption, marks a significant milestone in historical research and artificial intelligence. The project has already achieved breakthroughs, utilizing AI to scan and interpret paragraphs from the scrolls found in Herculaneum. 

Read more


OpenAI Announces First Partnership With a University

OpenAI recently announced a partnership with Arizona State University (ASU), providing full access to ChatGPT Enterprise for coursework, tutoring, and research. ASU plans to leverage ChatGPT for personalized AI tutoring in STEM subjects and to develop AI avatars for creative study assistance. The collaboration aims to enhance student learning experiences while ensuring privacy and intellectual property protection, with ASU exploring innovative applications of ChatGPT in higher education.

Read more



OpenAI Forms a New Team To Study Child Safety

OpenAI has established a new Child Safety team to address concerns about the potential misuse of its AI tools by children, reflecting a broader industry trend of tech companies dedicating resources to comply with regulations such as the U.S. Children’s Online Privacy Protection Rule. The team's role includes managing processes, incidents, and reviews related to underage users, and OpenAI is actively seeking a child safety enforcement specialist to join the team. This move follows OpenAI's collaboration with Common Sense Media on kid-friendly AI guidelines and highlights the company's commitment to ensuring responsible AI usage among minors amid growing concerns about the risks and benefits of children interacting with AI tools.

Read more

Semafor Reporters Are Going To Curate the News With AI

Semafor, a newly revamped politics site, is introducing Signals, a product that aggregates and summarizes news stories from various sources on daily topics. This curation process is aided by MISO, an AI-powered search tool that helps reporters find relevant content, particularly from non-English sites, making the process more efficient. MISO has been dubbed as “a custom bot built on OpenAI’s platform and using Microsoft’s Bing search engine.” While MISO conducts searches in multiple languages, the final curation and editorial judgment are still performed by humans. Semafor's collaboration with Microsoft aims to explore responsible AI usage in journalism, highlighting the evolving role of technology in news reporting.

Read more


Check out how GIANT Remixes are helping GIANT kids to fuel their creativity with AI as a tool with our latest GIANT Postcards: Kawaii Love Monster Valentine’s Day Postcards. You can even have the chance to create your own!

Janet Bloom