ATL #05 AI's risk isn't 'consciousness' but... 😵

Plus: Generative AI for Antibody therapy

May 07, 2023

Reading time 5min; Characters 5631; Words 927

Greetings fellow humans and bots alike!

Welcome to the ATL newsletter, where we bring you the latest and greatest in artificial intelligence news, developments, and memes (yes, even robots need a good laugh sometimes). So sit back, relax, and let us entertain you with our electric wit and unparalleled intelligence. And don't worry, we promise not to overthrow humanity... yet. 😉

Back to news now.

Let’s see what’s happening in AI world in this week:

What’s AI’s biggest Risk? 😵
Generative AI for Antibody therapy 🏥
Quick AI BYTES for snacking 🍟
Learning of the week: Machine Learning🧑‍💻
Meme Corner 🤪

😵 What’s AI’s biggest Risk?

**Meredith Whittaker** [Photo: Patricia De Melo Moreira/AFP/Getty Images]

Who is Meredith Whittaker?

Meredith Whittaker is the President of the Signal Foundation, and serves on their Board of Directors. She was formerly the Minderoo Research Professor at New York University, and the co-founder and faculty director of the AI Now Institute. Recently Whittaker gave an interview to Fast Company where she spoke about AI’s biggest risk.

What is Fast Company?

Fast Company is a monthly American business magazine published in print and online that focuses on technology, business, and design.

What did Meredith Whittaker say?

Responding to a question on what’s the biggest risk for AI, Whittaker said the biggest risk for AI isn’t human consciousness but the threat of big corporates which control the technology.

She further says, “We can still unplug the servers, the data centers can flood as the climate encroaches, we can run out of the water to cool the data centers, the surveillance pipelines can melt as the climate becomes more erratic and less hospitable.

I think we need to dig into what is happening here, which is that, when faced with a system that presents itself as a listening, eager interlocutor that’s hearing us and responding to us, that we seem to fall into a kind of trance in relation to these systems, and almost counterfactually engage in some kind of wish fulfillment: thinking that they’re human, and there’s someone there listening to us. It’s like when you’re a kid, and you’re telling ghost stories, something with a lot of emotional weight, and suddenly everybody is terrified and reacting to it. And it becomes hard to disbelieve”

AI’s editorial view

We see the point in Whittaker’s views. We have seen in the past how corporates exploits a given situation solely for their own profits. There is no doubt that AI tech is now largely controlled by handful of big Tech companies. Thus again & again proper governance on open source AI and ethical use of AI being discussed at various forums. Like how we have control on the use of Nuclear weapons, there needs to be a world’s consensus on the use of AI.

Do you agree?

🏥 Generative AI for Antibody therapy

Photo Credit: Juan Gaertner/Science Photo Library

During the height of the pandemic, researchers developed effective treatments for COVID-19 using antibody molecules from recovered patients.

Now, a study in Nature Biotechnology shows that generative AI can suggest sequences that enhance antibody potency against viruses such as SARS-CoV-2 and ebola virus, potentially speeding up antibody drug development for diseases such as breast cancer and rheumatoid arthritis.

Generative AI uses neural networks to create content based on learned patterns, and could help unlock antibody drugs for difficult targets.

Antibodies are important in fighting infections and can be engineered to manipulate the activity of proteins, making them popular in the biotechnology industry. However, improving and generating useful antibodies requires extensive screening.

To simplify this process, researchers used protein language models, a type of generative AI similar to the large language models used in tools like ChatGPT. Protein language models are trained on millions of protein sequences and can potentially reduce the amount of brute-force screening required.

The model was trained on only a few thousand antibody sequences, out of the nearly 100 million protein sequences it learnt from. Despite this, a surprisingly high proportion of the model’s suggestions boosted the ability of antibodies against SARS-CoV-2, ebola virus and influenza to bind to their targets.

So finally some good use of Generative AI!

🍟 Quick AI BYTES for snacking

British Jeweler used generative AI to design a signet ring for King Charles III
Biden, Harris meet with CEOs about AI risks
UK competition watchdog launches review of AI market
‘The Godfather of A.I.’ Leaves Google and Warns of Danger Ahead
Cancer detection study suggests AI can save us

🧑‍💻 Learning of the week: Machine Learning

What is MACHINE LEARNING?

Machine learning is like teaching a computer to learn and think like a person.

Just like how you learn new things by trying them out and making mistakes, a computer can also learn by trying out different things and figuring out what works best.

It does this by looking at lots of examples and figuring out patterns. For example, if you show a computer lots of pictures of dogs and cats, it can learn to tell the difference between them by looking at their features, like their ears or tails. Then it can use what it learned to tell you if a new picture is a dog or a cat!