How can we develop AI systems that are more respectful, ethical, and sustainable? - Highlights - DR. SASHA LUCCIONI

How can we develop AI systems that are more respectful, ethical, and sustainable? - Highlights - DR. SASHA LUCCIONI

What are the pros and cons of AI’s integration into our institutions, political systems, culture, and society? How can we develop AI systems that are more respectful, ethical, and sustainable?

Dr. Sasha Luccioni is a leading scientist at the nexus of artificial intelligence, ethics, and sustainability, with a Ph.D. in AI and a decade of research and industry expertise. She spearheads research, consults, and utilizes capacity-building to elevate the sustainability of AI systems. As a founding member of Climate Change AI (CCAI) and a board member of Women in Machine Learning (WiML), Sasha is passionate about catalyzing impactful change, organizing events, and serving as a mentor to under-represented minorities within the AI community. She is an AI Researcher & Climate Lead at Hugging Face, an open-source hub for machine learning and natural language processing.

THE CREATIVE PROCESS · ONE PLANET PODCAST

Data centers are crucial for powering AI and they already account for up to 1.5% of global electricity use, according to the International Energy Agency. This will significantly rise with the rapid growth of artificial intelligence. On the other hand, AI might assist us with climate change mitigation.

SASHA LUCCIONI

The way I got into this field was working on the environmentally beneficial applications of AI, and I do believe that that's an impactful way of using AI techniques because there's so much data about the climate, satellite data, and sensor data, and the way to go about this is to work with domain experts. AI is never going to solve the problem on its own, but it can be a tool. So I think that there's a lot of promise there.

*

My TED Talk and work are really about figuring out how, right now, AI is using resources like energy and emitting greenhouse gases and how it's using our data without our consent. I feel that if we develop AI systems that are more respectful, ethical, and sustainable, we can help future generations so that AI will be less of a risk to society. 

And so really, artificial intelligence is not artificial. It's human intelligence that was memorized by the model that was kind of hoovered up, absorbed by these AI models. And now it's getting regurgitated back at us. And we're like, wow, ChatGPT is so smart! But how many thousands of human hours were needed in order to make ChatGPT so smart?

Understanding the Current State of AI

When people say, oh, yeah, AI is going to help everyone or change humanity or all these claims, they don't realize people don't have access to the Internet in some places or cell phones or the fact that the data used by AI models is not representative of many parts of the world. It's mostly in English, and the data generated on the Internet and AI is mostly by educated, white, male users who post on forums. So there are whole generations and whole regions of the world that are not represented in this data. And so I think that all these claims of the universality of AI or how it's going to help everyone are techno-optimistic. 

The Efficiency of the Human Brain vs AI Models 

I think it's kind of important to think about the fact that originally, neural networks were inspired by neurons, really by our brains, but fundamentally, brains are really different than AI models, currently. Brains are quite efficient. For example, we won't be using all of our neurons all the time. If we're recognizing a face, or if we're speaking, or if we're drawing, or if we're watching a movie, or reading a book, there are different zones. And I think that that's kind of the core of the issue because whereas we can kind of use a portion of our brain, depending on the activity, neural networks don't have that built-in efficiency.

The Challenges of AI Governance and Legislation

The US Executive Order on AI still does need a lot of operationalization by different parts of the government. Especially, with the EU and their AI Act, we have this signal that's top down, but now people have to figure out how we legislate, enforce, measure, and evaluate? So, there are a lot of problems that haven't been solved because we don't have standards or legal precedent for AI. So I think that we're really in this kind of intermediate phase and scrambling to try to figure out how to put this into action. 

The Importance of Critical Thinking

I think it's really important to stay, once again, skeptical of AI, but also learn about it and see it not as some magical thing, but more as a technology. A technology that works but also doesn't work. A technology that comes with costs and benefits. So I think it's really important to stay aware of this change that is happening and not just take it as a given, essentially, because young people are using AI almost all the time. It's almost integrated, it's part of their lives nowadays, and maybe they don't stop and think about what is this data that I'm sending to TikTok or Snapchat, like what is happening? This filter, does it work as well for me as for a friend of mine that's either female or has darker skin or is wearing glasses? Things like that. I think it's really important, and it's something that I try to do with my own kids. That kind of critical thinking is going to be more and more important. 

This interview was conducted by Mia Funk with the participation of collaborating universities and students. Associate Interviews Producers on this episode were Sophie Garnier and Sebastian Classen. One Planet Podcast & The Creative Process is produced by Mia Funk. Additional production support by Katie Foster.

Mia Funk is an artist, interviewer and founder of The Creative Process & One Planet Podcast (Conversations about Climate Change & Environmental Solutions).

 
logo-white-space interviews podcasts.jpg

 
 

CLICK FOR MORE

PODCAST INTERVIEWS