Intro
Welcome to the AitopiA Initiative page, great that you have found us, and great to have you here!
We, Natasha and Mark, have started this initiative because we are both involved in the AI ecosystem and because we care about our planet and our human society.
AI offers tremendous possibilities to improve our lives. It can automate away the boring stuff, it can make sure that security works in real-time, and it can help doctors detect diseases. However, this power comes has to come with great responsibility. AI is a powerful tool, and we need to make sure that it is used for the betterment of society.
Current AI models (and the engineers creating the models) have a range of challenges laid out for them when it comes to responsibility, ethics, and sustainability. Think about predictive models that discriminate on gender, age, and race and the amount of power that these models hold over society (frequently without the society knowing about it).
Moreover, for the past years, most of the trends within AI research are directed towards marginal improvements in model performance, which results in more and more power-hungry algorithms. As we saw somewhere on the Internet:
"Your machine learning model might be melting icebergs!"
We believe that much more attention has to be given to the power consumption and efficiency of the models and that the (lack of) environmental impact has to be one of the driving metrics in developing new state-of-the-art AI algorithms.
This is where AitopiA comes in: We want to create a platform where we can share information on and discuss AI that is responsible, sustainable, and ethical. This is not only a challenge for the industry, but also for the AI community. We want to help to make this happen.
Videos
On this page, we will place recordings of our previous events such as webinars and meetups. Also, we might highlight some videos on responsible and sustainable AI which we really find insightful!
Webinars & Meetups
The first webinar we organized on the topic of Sustainable and Climate AI was in collaboration with Xomnia. To kick things off, we invited Angela Fan and a duo of Peter van de Putten and Jeroen van der Most.
Angela Fan talked about the technical side of things: how can we create more sustainable AI models without sacrificing too much. Angela is a PhD student at INRIA Nancy and Facebook AI Research Paris and is an active contributor and thinker in the field, as she recently co-organised the SustaiNLP workshop at the EMNLP 2020 conference.
Following Angela’s talk, Peter van de Putten and Jeroen van der Most showed us a more artistic way to approach the problem and raise awareness around climate change by letting the AI speak for itself by generating text through an AI model.
The second webinar we organized on the topic of Sustainable and Climate AI was also in collaboration with Xomnia. In this edition, we invited Lucia Loher, the Product Lead at Jina AI and Ricardo Vinuesa, who is an Associate Professor and Affiliated Researcher in the AI Sustainability Center at KTH Royal Institute of Technology.
Why does Jina AI contribute to the sustainable AI goal? In our opinion, democratization of AI is an important step in making the usage more sustainable. By sharing the pre-trained models, and making the existing algorithms as efficient as possible out-of-the-box, it becomes less necessary for each individual developer to train and fine-tune their own models!
As for Ricardo's work, his paper on the role of artificial intelligence in achieving the UN Sustainable Development Goals was published in Nature Communications Journal and inspires great conversations on the idea that AI can both enable and inhibit the accomplishment of the set sustainable development goals.
Blog
Coming soon!
Contact
Coming soon!
