A week in Generative AI: Lilli, Biomes and a copyright ruling
News for the week ending 20th August 2023

Thereās lots to unpack this week with some really good long reads as well as news on AI copyright, and from Google, McKinsey and OpenAI.
These Women Tried to Warn Us About AI
I shared this on LinkedIn earlier in the week, but itās too important an article not to share again here. A must read for anyone interested in the ethics of artificial intelligence but more importantly ethics and equality in general.
Judge Rules Wholly AI-Created Art Can't Get Copyright Protections
This is the first significant ruling Iāve seen in an AI copyright case, essentially ruling that copyright is only applicable for human-made content. The judge said that copyright can only be granted to āan originator with the capacity for intellectual, creative, or artistic laborā. Fully expect this to change in the coming months/years!
Google DeepMindās CEO Says Its Next Algorithm Will Eclipse ChatGPT
Great article from Wired on Gemini, the next big generative AI model thatās coming from Googleās DeepMind. Iāve been speculating for a while that they would combine the experience they have from AlphaGo and strategy/long-term planning with the capabilities of a large language model to leapfrog ChatGPT. Really excited to see what they release later this year!
McKinsey launch Lilli, their generative AI tool
Lilli is a platform that provides McKinsey employees with a āstreamlined, impartial search and synthesis of the firmās vast stores of knowledgeā. I suspect this is built using their partnership with Cohere, and is the first of these types of models Iāve heard of being deployed at scale.
OpenAI acquires AI design studio Global Illumination
This is the first acquisition OpenAI has ever made, so itās an interesting one on many levels. It seems the thing Global Illumination are best known for is for creating Biomes which is kind of like an open source Minecraft. Be interesting to see where this one goes!
Googleās AI Can Now Summarise an Article Directly in Search Results
Iām not sure how useful this feature will be in practice, but itās good to see Google experimenting with how they can enhance the search experience with generative AI.
This weekās long reads
AI Supremacy - How far are we from AGI?
Every.to - The internet is still early
āThe future is already here, itās just not evenly distributed.ā
William Gibson