🤔Joe Tsai: 99% of Those Discussing AI Are Men. What Gender Should the Smartest AI Be?

Описание: Hey, folks! There's a super thought - provoking topic hitting the tech scene once again. Joe Tsai, the co - founder of Alibaba, has tossed out an incredibly interesting idea. He points out that in the current AI discussion frenzy, a staggering 99% of the participants are men. This really makes us pause and wonder: in this AI discussion arena dominated by male voices, what gender should the smartest AI be designed as?​
The conversation about AI's gender isn't new at all. When we look at practical applications, many of our everyday intelligent voice assistants, like Tmall Genie, are programmed with female voices. Jack Ma once mentioned that at Alibaba, while female technical staff only make up 13%, a whopping 51% of the designers and nearly half of the product testers are women. Women clearly have a unique say in setting the standards for beauty and quality. From this angle, it seems that an AI with some "feminine traits" can offer a more warm and detailed experience.​
However, when we dig into the research data, things get a bit more complex. Researchers at the University of Michigan monitored 2,457 questions across three models, namely Flan - T5, LLaMA2, and OPT - instruct. They discovered that when handling legal - related tasks, neutral or male - role - based operations are more efficient. Plus, AI systems inherently show a preference for male - role and gender - neutral roles. Moreover, the "Research Report on Promoting Gender Equality in AI Algorithms", released in 2021 with the support of the United Nations Entity for Gender Equality and the Empowerment of Women, revealed that AI algorithms exhibit gender discrimination in numerous application scenarios. For instance, an AI face - recognition platform once misidentified a picture of "a man holding a fruit basket" as "female", but when just the head was cropped and analyzed, it was correctly recognized as "male". Amazon's automated recruitment algorithm also used to give low scores to resumes with the word "female" in the keywords. This indicates that during the training of AI algorithms, due to the influence of gender biases in human society, women seem to be at a disadvantage in the AI realm.​
Digging deeper, the gender discrimination in algorithms mainly comes from the fact that the datasets used to train AI mirror the gender biases in human society. Algorithm engineers often lack sufficient awareness of this issue and fail to include solutions for gender biases in development requirements. As a result, these biases are amplified through the algorithms. Looking at the composition of practitioners, among those surveyed, in 72% of the AI algorithm projects they were involved in, the proportion of female developers was less than 30%. And in 66% of the teams, the proportion of female decision - makers was less than 30%. If women's voices are absent during the AI research and development process, will the "personality" shaping of AI end up being lopsided?​
The question Joe Tsai posed truly deserves our careful thought. Should we go along with the current situation where AI leans more towards a male - centered approach in terms of algorithms and performance? Or should we actively work to break these biases and enable AI to have a more balanced set of "gender characteristics", incorporating more of women's sensitivity, creativity, and finesse? What are your thoughts? Jump into the comment section below and let's chat!​
#Alibaba #Vlog #JoeTsai #Alibaba #AI #AGI #Gender