The importance of being active in research and ethics stays in understanding both. Our guest, Giada Pistilli, works as Principal Ethics at Hugging Face, one of the most remarkable companies of the moment. Read about her experience!
The importance of being active in research and ethics stays in understanding both. Our guest, Giada Pistilli, works as Principal Ethics at Hugging Face, one of the most remarkable companies of the moment. Let’s learn from her experience from philosopher to ethicist with a deep dive into the STEM world. Ester Liquori, from WAI Italy, hosted the interview and discussed topics like inclusion and women's empowerment with Giada.
Tell us about your role for one of the most exciting AI companies of the moment.
I am a philosopher by training, but I have been working alongside Machine Learning engineers and designers for almost three years in R&D projects linked to ethics and AI.
I started my role as an ethicist at Hugging Face last May, and my missions is to help shape the company's ethical endeavors and drive the narrative of the role of ethics in Machine Learning development and deployment.
On a daily basis, this translates into working with decentralized teams across the company (e.g., the Multimodal project's ethical charter) and in the widespread of the Machine Learning community to foster ethical decision-making grounded in theory and practice (e.g., my op-ed in the 'conscious AI' debate).
Currently, I am coordinating the Hugging Face platform's content moderation and policy with policy and legal experts and working on several research projects for our BLOOM release with the science team. [note form the author: BLOOM is the NLP open-source model created by BigScience, a Hugging Face initiative]
How did your journey in STEM begin? What brought you to the field of ML/AI?
After two years of working and traveling after my bachelor's in political science and international relations, I realized I wanted to do a master's in political philosophy and ethics, my eternal love.
So I applied at Sorbonne Université, where I got to know ethics professors like Michel Puech, who taught us the importance of applied ethics to technology back in 2017 when AI ethics was just something existing "in thin air". I decided I wanted to mix my expertise in political science with some applied ethics, and that's how I ended up writing a master's thesis on AI at the service of the community.
There's where I learned to do empirical philosophical research and discovered the company that allowed me to continue my research with them with a Ph.D. industrial contract at the end of 2019.
Being a first-generation and coming from a very humble family, it was out of the question to do research and continue to be precarious. So in the end, the professional experience allowed me to be where I am today! It was and still is an exciting journey, and I love doing interdisciplinary research with stellar colleagues. I argue that is the future of AI ethics, where ethics (as in the philosophy tradition) serves as a lens to foresee the possible future and guide human actions while bridging the gaps with other scientific fields.
What is the most significant change you had observed in the field compared to when you entered the field early in your career?
I would say people seem to be very interested in ethical considerations linked to AI systems usage and AI ethics in general, compared to 5-6 years ago.
Users tend to be more aware, policymakers ask (usually) good questions, and the industry is (usually) willing to investigate AI ethics-related issues.
However, with the generalization and expansion of this scientific domain, which is in my eyes interdisciplinary with bases firmly anchored in philosophy, a very high percentage of the human investment in AI ethics is composed by:
Industries hiding behind a veneer of ethical reflection
Pure ethics washing.
Of course, there are always positives and negatives, and since AI is very fashionable today, it attires both; I would give advice always to have a base of operability and concrete actions.
What does diversity mean to you in the AI and Data Science world?
Having been fortunate enough to contribute with over a thousand researchers to the BigScience project, I can say that I have experienced true diversity! Cultural, religious, philosophical, professional diversity, and so on.
I think it is crucial that the AI community today is diverse in its complementarity, thus bringing out different visions while contributing to a common project.
This is by no means a simple goal since those who say "diversity" also say "communication diversity" (e.g., the word "value" means something to me as a philosopher, while to a mathematician, anything but!).
Diversity means being allowed to express themselves while having a mindset predisposed to accommodate potential conflict. All this is to say that diversity in the AI field is necessary and difficult to achieve, but not impossible!
How have you seen the role of women in AI evolve during your career, and what future do you see at the rate women pioneers are emerging in the AI/ML space?
To be honest, I've seen a lot of them! I am lucky enough to know extraordinary women in this field who specialize in computational linguistics, computer science, computer vision, Natural Language Processing, data science, statistics, mathematics, ML engineering, model evaluation, speech recognition, philosophy, and ethics of AI and I could go on! Nevertheless, it is definitely an environment that can still improve in welcoming women and other genders. But this battle is not just about AI and requires collective mobilization and participation at the social and political levels.
What do you think are the reasons why few women are interested in AI-related professions?
Perhaps because of the stigma that still exists today when people are interested in the so-called "hard sciences"?
To be honest, I see a lot of interest in AI from women, and not only on the social sciences side. In Hugging Face, my women colleagues are among the most reputable and hard-working professionals I have ever met!
In what ways do you think technology has served women today? In what ways does it help women's empowerment?
This is a difficult question that is likely to contradict my previous answer! Because even though so many women are involved in developing AI systems and everything around them, these new technologies have not historically been overtly helping women's empowerment.
I will not list the negative examples I am thinking of, preferring to focus on the future, which can be positive but still needs to include women in their design.
Technology is an instrument of power, depending on who holds it and the use they make of it. But they remain tools.
Therefore, I think the solutionism for empowering women lies more in human support than technological support, although it is important to continue pushing for as broad an inclusion as possible.
Thank you Giada Pistilli. We hope to see the effects of your essential contribution to your company's outputs to the world!
Do you want to contribute to building a better future for the AI and women's workspace?