AI & sustainability [Webinar]
AI can help save the planet, but a big AI model can also burn the co2 equal to almost 125 round-trip flights between New York and Beijing. At this webinar, we explore how we can increase the positive possibilities and impact of AI while lowering the planetary and economic downsides.
About
AI can help save the planet, but a big AI model can also burn the co2 equal to almost 125 round-trip flights between New York and Beijing. When AI and ML emits co2, it does so with the use of compute power, thereby burning both energy and money.
So how can we increase the positive possibilities and impact, while lowering the planetary and economic downsides?
First we dive into how AI and machine learning can be an extremely powerful tool in fighting climate change & saving money, to mention a few of the possibilities;
- Enabling use of more green energy, reducing energy use and impact on current electrical systems
- Optimizing cities, buildings & production for less energy, water & resource consumption
- Improving transportation energy use, by lowering activity, improving efficiency in supply chains, fuel use & driving
- Supporting sustainable behavior through footprint insights & nudging
- Food farming & forests, by aggrotech precision farming, forest management and monitoring
That being said, it comes at the cost of money, energy & co2, to decrease these negative impacts of AI, and increase the possible positive impact of AI and ML, the following must be considered;
- Small data rather than big data
- Simple explainable & more efficient models, rather than deep learning
- Local models, rather than global through federated learning
- Balance accuracy & efficiency, by using pre-trained models, and design for efficiency from the beginning
- Let human and machine reinforce each other
Finally let’s talk about how we can all take action today to save the planet, money & energy.
For more info see:
https://arxiv.org/pdf/1906.05433.pdf
https://www.climatechange.ai/about
https://www.technologyreview.com/2019/06/06/239031/training-a-single-ai-model-can-emit-as-much-carbon-as-five-cars-in-their-lifetimes/
https://www.computerweekly.com/feature/How-to-make-AI-greener-and-more-efficient
Speaker
Emilie Lundblad
Managing Director of Data & AI at Amesto Nextbridge
Speaker at Make Data Smart & Genesis Conference
12 Microsoft Data & AI certifications
Minor in Machine learning & AI from MIT (Guest lecturer fall 2022)
Writing master thesis in AI on Executive master of IT Management at ITU (Guest lecturer fall 2022)
Candidate in quantitative finance & econometrics
12+ years' experience with data & analytics, 6 of them in Data & AI consulting.
Startup entrepreneur & data nerd.
Join the webinar with a free trial membership
Not yet a member of Dansk IT? You can still join the webinar - here are your options:
Try a Dansk IT membership for free for 14 days. Sign up here: https://akademi.dit.dk/Bliv-medlem
It only takes 1 minute to sign up. With a trial membership you will gain access to the platform DIT akademi where you can join our many events and view video material from previous webinars.
Tilmeld
Login ind for at tilmelde dig.