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.
What’s going on with ChatGPT?
With ChatGPT dominating the headlines lately with the firing then rehiring of Sam Altman, the musical chairs played by the board, and the ominous bridge to AGI, Q* (pronounced Q-star), looming over it all, we wanted to stop and ask, How have YOU used ChatGPT since it was introduced to the public about a year ago? A quick survey in our GIANT community yielded some interesting use cases including creating a seven-child soccer carpool schedule (successfully!), checking third grade math homework (fail!), and writing a poem about The GIANT Room (undecided)!
Some other parents are turning to ChatGPT to plan birthday parties or assign age-appropriate chores. But, we would really love to hear about your ChatGPT experiences! Perhaps you’ve built your own GPT solution, and you’re ready to go head to head with OpenAI or Anthropic! OR, perhaps your child has encountered it in the classroom? Either way, please send us your use cases-whether useful, or useless, or downright avant garde! We will share our findings in a future GIANT AI Report. And, we will gladly share our ChatGPT-generated GIANT Room poem to anyone who participates!
Godmother of AI: Fei-Fei Li on Tech Development: “The Power Lies Within People” Many doubted her, but now Fei-Fei Li is a crucial leader and innovator in the world of technology when it comes to AI. The system she built years ago, “ImageNet,” is what many AI models today are based upon. These days, Li is an advocate for incorporating artificial intelligence into healthcare, and the co-founder of AI4ALL, an organization that aims to bring more diversity into the field. Hers is truly an underdog story.
OpenAI wants to work with organizations to build new AI training data sets In recognition of the inherent flaws in current AI model training data, which often reflects U.S.- and Western-centric perspectives and includes biases, OpenAI has launched Data Partnerships, aiming to collaborate with external organizations to develop improved, diverse data sets for AI model training. The initiative seeks to address biases and toxic language present in existing data, with the goal of creating safer and more beneficial AI. OpenAI plans to collect large-scale data sets representing human society, spanning various modalities, and will work with partners to ensure privacy and inclusivity in the process.
A Prisoner’s Dilemma Shows AI’s Path to Human Cooperation New research reveals that ChatGPT's engine tends to cooperate more than humans but also tends to overestimate human collaboration. The study, based on the prisoner's dilemma game-theory problem, suggests valuable insights for deploying AI in real-world applications. The findings indicate potential applications beyond natural language processing, such as urban traffic management and energy consumption, where AI could guide cooperative decision-making for more efficient outcomes. The study emphasizes the need for transparency, ethical guidelines, and responsible deployment practices as AI models adopt aspects of human nature.
Google's AI-powered search feature goes global with a 120-country expansion Google's Search Generative Experience (SGE), initially introduced at I/O 2023, is now expanding to Search Labs users in 120 countries and territories. SGE, which provides generative AI summaries at the top of search results, now supports additional languages, including Spanish, Portuguese, Korean, and Indonesian. New features include an improved follow-up function for direct questions on the search results page, assistance in clarifying ambiguous translation terms, and the expansion of interactive definitions to coding and health-related searches.