Researcher inspires inclusion and change

Karima Kadaoui, remains in the United Arab Emirates (UAE) to utilize the power of artificial intelligence (AI) to assist disadvantaged communities and languages.

The Moroccan graduated with a master’s in machine learning as part of the university’s inaugural Class of 2022 and has made the seamless transition from student to research assistant.

As the world’s first graduate, research AI university, research plays a significant part in the curriculum and ethos of the institute to save real-world problems using computer science, computer vision, machine learning, natural language processing (NLP), and robotics.

With a focus on developing a diverse community of AI decision-makers within the Emirates, MBZUAI is a champion for gender equity in STEM with a 28% female student body (77 out of 272), five full-time female faculty members, 40 female researchers, and a gender balanced workforce (52% overall female workforce).

This year’s theme for International Women’s Day 2024 (March 8) is to ‘inspire inclusion’ and Kadaoui and her research certainly sends a powerful message as to how women in AI can play an important role.

What are your specific areas of research at MBZUAI?

My master’s thesis was on impaired speech recognition, which aims to learn the patterns of disfluencies that someone with a speech disability might have and convert those back to how they would sound if spoken by a healthy individual. A lot of people with speech impediments have additional motor disabilities, due to which they require assistance from their environment. Communication is therefore especially crucial for them. Most recently as a research assistant, I have been working on democratizing different NLP tasks on the Arabic language and its dialects, whether in the text or speech modalities. While current translation and speech recognition systems work fairly well on standard Arabic, it is different for spoken dialects, which can be almost unintelligible from country-to-country. If we aim to be inclusive, it is imperative that we develop systems that work on dialects spoken in real life.

Why do you feel it’s important to have a diverse and inclusive environment for women in AI?

I believe that innovation requires diversity, whether we’re talking ethnicity, gender, age, etc. Women are one side of one of the most important axes in this equation. Additionally, bias is not unheard of when it comes to AI systems. Including women is necessary for more objective systems and for matters or issues that a male-dominant environment might not notice (through no fault of their own, the same would apply if it were a women-dominant industry). Finally, since AI systems are used by both men and women, it only makes sense if the people behind building these systems accurately reflect the user distribution.

What are your longer-term career aspirations?

I aim to enroll in a Ph.D. to further deepen my knowledge of the field, and later join a research and development department in the industry. My preferred area remains leveraging AI for disatvantaged communities, and I hope to gain more experience with different modalities – such as NLP and computer vision – for a more holistic approach to assistive and inclusive AI. Maybe one day I’ll have my own company centered around the intersection of AI and social good.

How do you think AI research will benefit society?

AI can bring to life solutions to problems that were not previously solvable. It’s also an extremely fast-paced industry that keeps revolutionizing our reality as we know it with much shorter cycles compared to other research fields. It has the potential to improve healthcare, environmental sustainability, education, and public safety to name a few. That being said, harder efforts are needed to regulate AI properly to be able to maximize its benefit to society and minimize its harm.