Noriko H. Arai (Japanese: 新井紀子, born October 22, 1962) is a Japanese mathematician, computer scientist, and researcher in mathematical logic and artificial intelligence.^{[15,42,43,44,72,77]} Born in Kodaira, Tokyo, Japan, she studied law at Hitotsubashi University, earned a Bachelor of Science in Mathematics magna cum laude from the University of Illinois Urbana-Champaign, and completed her Ph.D. at the Tokyo Institute of Technology. She is a professor in the Information and Society Research Division at the National Institute of Informatics (NII), where she also serves as Director and Professor of the Research Center for Community Knowledge (Center for Social Shared Knowledge), as well as a professor in the School of Multidisciplinary Sciences (Department of Statistical Science) at The Graduate University for Advanced Studies.^[46,72]

Her research focuses on mathematical logic, including propositional logic and proof theory, and extends to artificial intelligence, natural language processing, educational technology, and distance education.^[43,44] She is widely known for leading the Todai Robot Project, an initiative to develop artificial intelligence and robots capable of passing the entrance examinations for the University of Tokyo.^[15,72,77] The "Todai" AI ranked in the top 1% of Japanese students in mathematics and achieved a 70% passing score on the entrance exam, prompting Arai to express concern about education because the program could pass exams that human students fail without possessing an understanding of concepts or ideas.^[40] In addition, she is known as the founder, developer, and operator of Researchmap, the largest database and social network for researchers in Japan.^[45] Arai is also recognized for her general-audience books and lectures addressing the impact of AI research on educational issues, and was one of 15 artificial intelligence researchers invited by French President Emmanuel Macron to attend the presentation of a major French artificial intelligence research initiative.^[41,45]

Professor Noriko H. Arai was born in Tokyo.^[1] During her time at Tokyo Metropolitan Kunitachi High School, her strongest subject was Japanese language arts, which led her to pursue a liberal arts undergraduate degree in the Faculty of Law at Hitotsubashi University, where she enrolled in 1981 according to Japanese sources or in 1982 according to Korean sources.^[1] At Hitotsubashi University, where mathematics was a required subject, Arai developed an interest in understanding the process behind mathematics, leading her to take the SAT and TOEFL exams and study abroad in the Department of Mathematics at the University of Illinois Urbana-Champaign. In 1985, she graduated magna cum laude with a bachelor's degree in mathematics from the University of Illinois Urbana-Champaign.^[74] In 1984, she entered the five-year integrated doctoral program in mathematics at the University of Illinois Graduate School on a scholarship, marrying mathematician Toshiyasu Arai while at the university.^[3] Regarding her graduate degree from the University of Illinois, English and Japanese sources report she completed a master's degree in 1990 instead of the doctoral program, while Korean sources state she received a master's degree in 1987 and a doctorate in mathematics in 1990.^[3] Returning to Tokyo in 1990 while seven months pregnant, she received support from the university and finished her degree from the Faculty of Law at Hitotsubashi University—a program emphasizing human resources rather than a J.D.—graduating in March 1995 according to English and Japanese sources, or in 1994 according to Korean sources.^[1]

In 1994, she joined the Faculty of Information Sciences at Hiroshima City University as a research assistant/associate, though English sources state she became an assistant professor in the Faculty of Information Science at Hiroshima University where her husband worked.^[4] She received her doctorate in science from the Tokyo Institute of Technology with a dissertation titled "On Lengths of Proofs in Propositional Calculi," completing the degree in March 1997 according to English, Japanese, and Korean records, or in 1998 according to German sources.^[5,6] She was a visiting researcher at the Fields Institute in 1998, and in 1999 worked as a visiting member at the Institute for Advanced Study in Princeton as well as a visiting researcher in the Faculty of Information Studies at the University of Toronto.

Arai joined the National Institute of Informatics (NII) in 2001, appointed as an associate professor in the Principles of Informatics Research Division according to German, Japanese, and Korean sources, or as a professor according to English and French sources.^[73] In 2002, she was appointed associate professor in the Department of Informatics at The Graduate University for Advanced Studies (SOKENDAI) and served as a part-time lecturer. She became a part-time lecturer for liberal arts education at Hitotsubashi University in 2004, followed in 2005 by appointments as an adjunct or affiliate associate professor at the Tokyo Institute of Technology and a visiting lecturer at the University of Tokyo's Institute of Social Science. In 2006, she became a professor in the Information and Society Research Division at NII, a professor at SOKENDAI, and an adjunct or affiliate professor at the Tokyo Institute of Technology. She became director of the Research Center for Community Knowledge (Social Shared Knowledge) at NII in 2008, and in 2009 served as the Osmer (or Ottmar) Distinguished Visiting Professor of Natural Sciences at International Christian University. She established the Institute for Research in Better Education (Research Institute of Science for Education) in 2017, serving as its representative director, and later became an advisor to Toyota Systems Corporation in 2019 and a member of the Tokyo Metropolitan Board of Education in 2021.

At NII, Arai leads multiple projects, including founding Researchmap, Japan's largest social network for researchers with profiles for over 250,000 researchers. She also leads the NetCommons project, an open-source content management system used by more than 5,000 Japanese schools for their homepages and groupware, making it the most popular CMS in Japanese schools. In artificial intelligence research, she has led the "Reading Skill Test Project" since 2011 to measure basic reading skills for young people, and since 2011 has served as director of the Todai Robot Project, developing the AI program "Todai" with the goal of passing the University of Tokyo's two-stage entrance examination by 2021. The Todai robot achieved high scores in mathematics and English, outperforming 80% of university applicants, but struggled on questions requiring an understanding of values. At the 2017 TED conference, she stated that the program's success demonstrates human education should focus more on problem-solving and creativity rather than rote learning. Arai has served on program committees for many international conferences in her research fields, was awarded the Nexplo Award in 2016, and in March 2018 was among 15 leading artificial intelligence researchers invited by French President Emmanuel Macron to announce a major French initiative for AI research.

Arai's early publications examine mathematical proof systems, particularly within the fields of calculus and combinatorics, while her more recent research has focused on the intersection between artificial intelligence and education.^[1,11] Alongside several co-authors, she was the last author on a 2018 paper titled "Can an A.I. win a medal in the mathematical olympiad? – Benchmarking mechanized mathematics on pre-university problems" published in AI Communications, which created a problem library designed to support the integration of mechanized mathematics and natural language processing toward end-to-end automatic mathematical problem solving.^[7] This work extended Matsuzaki et al. (2017), "Semantic parsing of pre-university math problems," in which Arai and her team developed an end-to-end math problem solving system that accepts natural language input.^[8]

In 2011, the National Institute of Informatics launched the Todai Robot Project ("Can a robot get into the University of Tokyo?"), an artificial intelligence initiative aiming to create an AI robot, named "Torobo-kun," capable of passing the University of Tokyo entrance examination by 2021.^{[13,14,12,72]} Arai became the director of this research and development project in 2011.^{[13,14,12,71]} The selective examination includes eight subject tests across sciences, social studies, Japanese and Chinese classics, English, and mathematics, with fewer than 15,000 of the over 500,000 annual high school exam participants permitted to advance to the test.^[13,14,12] In 2015, Arai published "The impact of AI: can a robot get into the University of Tokyo?", announcing that the robot achieved a score above the median human test taker, placing in the top 20% of examinees with an overall standard score of 57.1 across eight subjects in five fields, scoring better than 80% of applicants and achieving marks competent enough to pass the entrance exams of 472 out of 581 private universities in Japan.^{[13,14,12,16,57]} However, the score remained below the average of admitted students and was not a passing mark, revealing low reading comprehension with a standard score of 49.7 in Japanese and highlighting that the AI could not understand context or read question texts.^{[13,14,12,57]} Because current AI learns statistically from big data without understanding meaning, it was determined impossible for it to acquire sufficient reading comprehension to pass the entrance examination, leading to the development of Torobo-kun being frozen in 2016. The project garnered international attention and was featured in Arai's 2017 TED Talk, where she argued that its results demonstrate that human education must place more emphasis on problem solving and creativity rather than rote memorization and repetition.^[17,75] Arai received several awards for her work on the Todai Robot Project, including the Association for Natural Language Processing's 30th Anniversary Paper Award in 2024.^[17,1]

While developing Torobo-kun, Arai observed that as many as 80% of human examinees scored lower than an AI unable to understand context, and a 2015 survey revealed that many junior and senior high school students similarly solved problems via calculation and memorization without understanding context.^[58] Recognizing that humans cannot compete with advancing AI through calculation and memorization alone, improving the reading comprehension and contextual understanding of students came to be seen as a more pressing challenge than improving AI capabilities, prompting the establishment of the Research Institute for Science in Education in 2016.^[59,60] Arai developed the Reading Skill Test to investigate reading comprehension and compare human skills with AI; created in collaboration with the Japanese Ministry of Education, Culture, Sports, Science, and Technology, the test has been administered to over 40,000 students.^[9,10,30,60] Arai is also an award-winning author whose book, AI vs. Children Who Can't Read Textbooks, addresses issues in Japan's education system and the future of AI.^[17] The book won the Okawa Publishing Award in 2018, a TOPPOINT Award for the First Half of 2018, the 66th Japan Essayist Club Award, and the 2019 Business Book Award, among others.^[17,23,27]

Arai developed NetCommons starting in 2001 and developed Researchmap starting in 2009, serving as the founder of Researchmap, the largest social network for researchers in Japan.^[76] In March 2018, she was one of 15 leading artificial intelligence researchers invited by French President Emmanuel Macron to join him for the announcement of a major new French initiative for artificial intelligence research.

In 2009, she received the NISTEP Award from the National Institute of Science and Technology Policy. In 2010, she was awarded the Ministry of Education, Culture, Sports, Science and Technology Prizes for Science and Technology, the Commendation for Science and Technology. She received the Trivia Publishing Award in 2015 and the Natural Language Processing Association's Outstanding Paper Award in 2016.

In 2018, she earned multiple recognitions, including the Ishibashi Tanzan Award, the Okawa Publishing Award, and the Avon Woman of the Year Award in Education. Also in 2018, she received the Japan Human Resources Department HR Awards Book Category Excellence Award, the PHP Institute's 27th Yamamoto Shichihei Award, the 66th Japan Essayist Club Award, and the TOPPOINT Award.

She was awarded the Business Book Awards Grand Prize in 2019. In 2022, she received the Minister of Education, Culture, Sports, Science and Technology Award for Science and Technology (Science and Technology Promotion Division) as well as the 4th Japan Treasure Summit Academic Award. In 2024, she received the Association for Natural Language Processing's 30th Anniversary Paper Award.

Her publications include Detour Pan! Can a Cellobot get into Tokyo University?, which was published by East Press in 2014. In 2018, Toyo Keizai published Raising children who can compete with AI and AI vs. Children Who Can't Read Textbooks. Detour Pan! Introduction to Mathematics for Survival (New and Expanded Edition) was published by Shinyosha in 2021. In 2025, Toyo Keizai published both Noriko Arai's Reading Comprehension Training and Shin Reading Comprehension.