AI researcher urges focus on efficiency, accessibility, and real-world impact beyond benchmarks
NEW BRUNSWICK, NJ / ACCESS Newswire / January 5, 2026 / Artificial intelligence continues to grow in size, cost, and influence. But according to Jia Xu, a long-time researcher in natural language processing, the true measure of AI success should not be scale alone. It should be a social good.
Drawing on over a decade of academic and industry experience across Europe, Asia, and the United States, Xu is raising awareness about the need to evaluate AI systems by their real-world impact, accessibility, and long-term value to society.
“Success isn’t just about personal achievements or recognition,” Xu says. “It’s about creating a meaningful impact that constructively changes the world.”
Why Social Good in AI Matters Now
AI systems are increasingly embedded in everyday life, from education and healthcare to communication and decision support. According to recent industry estimates, training a single large language model can consume as much energy as hundreds of households use in a year. At the same time, many organisations and communities lack access to AI tools because of cost, infrastructure, or technical barriers.
These systems have a direct impact on real people, not just hypothetical users,” Xu explains. “A model may excel in benchmarks but still fall short of benefiting society if it is inaccessible, biased, or not environmentally responsible.
Research has also shown that smaller, more efficient models can often deliver comparable performance for specific tasks while using a fraction of the computational resources. This opens the door to wider adoption, lower environmental impact, and improved transparency.
A Career Shaped by Global Perspective
Xu’s perspective on AI impact has been shaped by her global career. She completed her bachelor’s and master’s degrees in Germany, earned her PhD at RWTH Aachen University, and later held faculty positions at Tsinghua University and the Chinese Academy of Sciences. She has also conducted research at Microsoft Research and IBM Watson.
Over the years, Xu has authored around 50 research papers, holds 12 patents and provisional patents, and contributed to 18 top-ranking teams in major AI competitions, including second place in the Amazon Alexa Prize Social Bot Challenge.
“Competitions and real-world systems taught me that accuracy alone is not enough,” Xu says. “In practice, reliability, efficiency, and trust often matter more.”
Redefining Success in AI Research
Xu believes the AI community must expand how success is defined and measured.
“I measure success using two standards,” she says. “One is my own judgement as a researcher. The other is social feedback. If an idea is recognised and helps make the world better, then it matters.”
She points to the evolution of machine translation as an example. Once considered nearly impossible, it is now a standard tool used by millions every day. That progress came not from rapid scaling alone, but from sustained research focused on usability and impact.
What Individuals Can Do
Jia Xu emphasises that advancing socially responsible AI is not limited to institutions or large organisations. Individuals also play an important role.
“Everyone involved in AI can ask better questions,” she says. “Who does this serve? Who might it exclude? What is its implication to society in the long term?”
She encourages researchers, students, and professionals to prioritise efficiency, fairness, and long-term benefit in their work and to remain open to continuous learning.
“In my field, every success brings new challenges and questions,” Xu adds. “That is what keeps progress meaningful.”
As AI continues to shape the global economy and daily life, Xu calls on individuals to reflect on how they define progress.
Learn about the systems you use. Question how success is measured. Support ideas and practices that prioritise lasting social value.
“True success in AI,” Xu says, “is not just about building smarter machines. It is about contributing something that lasts and makes the world better.”
Media Contact:
Jia Xu
[email protected]
http://www.jiaxustevens.com
SOURCE: Jia Xu
