Restacking: New Year, Same [A]I?
Hi everyone,
Happy New Year, and thank you for joining me on this week’s restack.
It’s going to be a short one this week, as I’ve read a lot of 2024 prediction round-ups that really are all saying the same thing. My take is that the AI ‘hype’ is in for some more time – governments will start putting up guardrails to begin regulating the technology, and there will be more wishful thinking on what the technology could do.
Here are two resources that I’ve enjoyed this week, plus a small announcement…
All AI systems are models
The first resource of 2024 will be, you’ve guessed, it, the latest post from Erik J Larson on DeepMind Takes on Protein Folding.
Larson takes on AlphaFold, a system by Google DeepMind that has perhaps been glamourized for its ability to bring forward breakthroughs in protein folding. This is one of the longest unresolved issues in biology, physics, and chemistry.
On reading his article, I was rightly reminded that the outcome of AlphaFold must still be seen as a prediction rather than an absolute truth. It’s by definition a limited, and not entirely truthful, guideline to reality. It is not a discovery. The etymology of science takes the “scio, scire”, Latin for “knowing”. A model is a tool; it is not knowledge.
“We’re addicted to computer modeling, calling it ‘AI.’ The models might provide some benefit in a larger research context, but we treat them as ends in themselves. Like human expert prediction, the models are always limited and frequently wrong, but we’re in denial.”
First ISO safety standard for engineering and managing AI systems
Secondly, there’s been BIG news in the real AI world. The ISO – International Organization for Standardization has adopted the world’s first ISO standards for engineering and managing AI systems! As the director of Public Policy & Government Affairs at Shopify, Kostas Rossoglou outlines, this standard will form the first blueprint for developing AI systems ethically.
The ISO/IEC 42001 standard covers:
Creating clear AI governance policies and procedures
Evaluating the societal, environmental, and individual impacts of AI systems
Implementing best practices for data management
Addressing diversity and inclusiveness in AI systems
Ongoing monitoring and auditing
Will this standard change the game for quality and information security? We’ll see. It’s surely going to build trust in the complex AI supply chain for those who adopt it like AWS, apparently!
Honest AI announcements: what’s coming in 2024
This year is already shaping up to be busy. I’m currently working on writing a technical guide to practical applications of large language models (LLMs) with code examples and translating The Ethics of AI into Italian.
Plus, I have a new book in the works…
This is going to be a sequel to The Ethics of AI and will tackle wider issues about what defines humanity in the age of artificial intelligence. It will seek to answer questions like the role of technology in a possibly ‘postcapitalist’ era and how we develop systems that promote long-term shared equity.
Meanwhile, I’m continuing to commit to sharing quality, practical content on the ethics of AI here and on LinkedIn. There is so much more to talk about and share. I’m now sharing technical explainers to start the week, from how LLMs work to what a TOIA is, and I consistently post every Thursday to bring you deep dives into more niche topics and ‘restacking’ other substacks and resources that I have enjoyed for their fresh takes, critical analysis, and reliable information.
Join the beta reading community
What do you want to read about? I’m lining up some incredible guests in 2024 to share their voices and platforms, ranging from the intersection of AI and spirituality and philosophy, as well as practical issues, like how to use AI to grow companies and drive efficiency.
One such topic I am particularly interested in is how modern-day entrepreneurs can use the advances of AI to build incredible businesses, whether to make revenue or have a better work-life balance.
So with that, what do you want to read about? Let me know in the comments or contact me at alberto.chierici@nyu.edu.
Great start to the new year!