• The AI Daily Brief
  • Posts
  • Ban AI? Australia Asks Citizens How Strict AI Regulations Should Be

Ban AI? Australia Asks Citizens How Strict AI Regulations Should Be

Plus an AI camera that doesn't have a lens

The AI Breakdown First Five - Thursday June 1, 2023

A note: the First Five experiment continues this week!

5. A Poetry-Writing, Doom-Prophesizing Robot

Ameca is claimed to be the world’s most advanced robot. This week, Ameca showed up at the International Conference on Robotics and Automation (ICRA) in London, delighting attendees by writing poetry in multiple languages. Ameca was also the subject of a 60 Minutes Australia feature.

4. Teaching LLMs to Use Tools

Researchers are exploring new ways to train LLMs to use tools to expand their capacities without requiring burdensome computational costs and data. A new paper introduces GPT4Tools, a method that uses an instruction-following dataset and Low-Rank Adaptation (LoRA) to enable open-source LLMs like LLaMA to solve problems like visual comprehension and image generation — i.e. things like interpreting an image based on a given text description or generating an image from a text prompt. As we plow towards a multimodal future, researchers think this methodology could support use cases ranging from customer service bots to accessibility tools to content creation.

Check out ChatGPT’s summary of the paper using the Xpapers plugin and ChatGPT’s new Share feature.

3. Japan’s AI Copyright Precedent

Earlier this year, Getty Images sued Stability AI, alleging that it used Getty’s database of 12 million images to train their AI models without permission or payment. New rules coming out of Japan would make Getty’s case null-in-void. According to (admittedly sparse English-language) reporting, Japan’s government believes that copyright concerns have held the country back in AI development, leading Japanese Minister of Education, Culture, Sports, Science, and Technology Keiko Nagaoka to say that copyright will not be enforced in AI training.

2. Australia Asks Citizens If It Should Ban AI

Australia introduced a voluntary ethics framework for AI all the way back in 2018, and now the country has announced two new research papers as well as an 8-week consultation period designed to get feedback from Australian citizens on what their regulatory approach to AI should be.

The spectrum of Australia’s potential regulatory interventions on AI

1. The Camera With No Lens

Paragraphica is the camera reimagined for an AI world. Instead of using a lens to capture an image, Paragraphica uses location data, describes the setting including time and weather, and then turns that into an AI image directly on the device. Is it useful? Meh! Is it fun? Absolutely!

Meme of the Day

The Latest AI Breakdown Episodes

That’s it for today! Please let us know how you like this new format.

How do you like the First Five Format?

Login or Subscribe to participate in polls.

Signing off from the future’s past - NLW