人工智能
采用时间：两到三年
在人工智能（AI）领域，正在利用计算机科学的进步来创建智能机器，以完成任务并以与人类非常相似的方式做出决策。为了实现这一目标，AI基于机器学习进行推理，通过暴露于海量数据集和自然语言处理，提供计算机进行决策和预测的能力。这有助于人类与机器交互的方式类似于与其他人交互的方式。这些能力正在推动医疗保健，金融服务和教育等行业的一系列发展。随着基础技术在教育领域的不断发展，人工智能有可能以更直观的方式回应和吸引学生，同时减轻教师繁琐的任务，从而增强在线学习，自适应学习软件和研究过程。一些报告预测，到2022年，人工智能技术在教育领域的市场增长率为43％。

Artificial Intelligence
Time-to-Adoption Horizon: Two to Three Years
In the field of artificial intelligence (AI), advancements in computer science are being leveraged to create intelligent machines that accomplish tasks and make decisions in ways that closely resemble those of humans. To achieve this, AI draws inferences based on machine learning, which informs a computer’s capacity to make decisions and predictions through exposure to massive data sets, and natural language processing. This helps humans interact with machines in ways similar to how they interact with other humans. These capabilities are driving a host of developments in industries such as health care, financial services, and education. As the underlying technologies continue to develop within the education sector, AI has the potential to enhance online learning, adaptive learning software, and research processes in ways that more intuitively respond to and engage with students while also relieving instructors of tedious tasks. Some reports forecast a 43 percent market growth for AI technologies in the education sector by 2022.
 
Overview
Since being featured in last year’s Horizon Report: 2017 Higher Education, AI has continued to make headlines in a variety of sectors, from Tesla’s self-driving cars to Apple’s newest facial-recognition software debuted in the iPhone X. Education leaders have had a wide range of reactions to AI’s impacts on teaching and learning strategies, with sentiments ranging from optimism about its potential to transform and democratize education to skepticism because of its role in automating teaching and reducing jobs.Further, a recent Northeastern University and Gallup study revealed that although only 22 percent of those with a postsecondary degree believed that their studies prepared them to work with AI systems, 77 percent of respondents think AI will positively impact their lives.
As AI continues to develop across sectors, students who become knowledgeable about AI and gain more experience working with it could have a competitive edge in the workforce. India and China have made notable commitments to advancing and integrating AI into education. New Delhi’s Bennett University, in partnership with several U.K. institutions, received a grant to begin large-scale adoption, training, and skilling in AI across 25 postsecondary institutions.
In China, the previous head of the country’s Google operations is working with the government on a five-year plan to develop a two-step process for increasing AI knowledge transfer. The plan starts by upskilling educators in AI techniques such as machine learning; those educators will then leverage their new expertise to inform students about AI and share best practices across the country.
While notable examples of AI are being implemented in the classroom, administrative tasks are also using it to streamline their processes. Institutions are improving teacher evaluations using AI-enabled chatbots to record, organize, and provide detailed feedback from students.

Georgia State was recognized for creating Pounce, a chatbot that helps incoming students navigate the complex application process, presenting a personalized checklist for completing financial aid and enrolling in courses.
AI is advancing areas other than teaching and learning as well, including campus safety and management. The University of Texas at Austin (UT), for example, is using AI systems to track, label, and analyze traffic patterns in efforts to increase safety measures for pedestrians and alleviate high traffic burdens.
 UT is also using data to develop self-adjusting irrigation systems to reduce water consumption and significantly cut costs.
Relevance for Teaching, Learning, or Creative Inquiry
AI is a useful tool for implementing today’s leading pedagogical trends, such as personalized learning, while also encompassing a variety of technology-based solutions, such as machine learning and open educational resources. Carnegie Learning and OpenStax have partnered to create an affordable learning solution for developmental math students. By leveraging Carnegie’s Mika, an online math course enabled by machine learning and AI, and OpenStax’s free online textbooks, the joint effort aims to increase math scores through personalized tutoring and real-time feedback while also reducing costs to postsecondary students.
 To better expose students to real-world uses of AI, universities are partnering with corporations to research and identify use cases for the technology. The University of Technology Sydney recently announced an ongoing project with a major bank’s insurance practice aimed at increasing customer satisfaction. The resulting OnePath system leverages years of data from behavior modeling, text mining, and natural language processing to understand and distinguish the most relevant policy questions.
AI is also changing how students and teachers interact with learning materials. The University of Michigan announced that students enrolled in statistics courses would be using the newly developed M-Write, which uses machine-learning algorithms to help improve and streamline the writing process. By using automated text analysis techniques that can match vocabulary and topics, the system helps students identify weakness in their analysis, speeds up the grading process, and alerts educators about which students need additional assistance.
  As another example, by using learning analytics, online learning can adapt to automatically fit students’ needs and provide interventions to deliver “just-right, just-in-time learning.” Oregon State University piloted adaptive courseware in eight high-enrollment courses to deliver personalized content to students who might not otherwise receive individual attention. The university’s goal is to increase retention rates in these classes by proactively helping students succeed.
Even as AI is increasingly used to help students and institutions make informed decisions, a body of literature has emerged that cautions against relying strictly on AI systems. For example, in terms of admissions, education leaders are concerned about the “gray area” in AI decisions—that is, AI systems cannot determine which college is best for every student because such decisions are not wholly fact-based, and relying on AI in all such situations could diminish diversity in institutions.
  However, AI is proving useful for completing time-consuming, tedious tasks, freeing instructors to focus on creating engaging learning experiences. A professor from Shenzhen University and Huazhong University of Science & Technology developed an AI-based framework for creating realistic textures that could further improve virtual worlds. By developing a process to automate these textures at a large scale, researchers can devote more time and resources to improving video game design, virtual reality, and animation.
Artificial Intelligence in Practice
The following links provide examples of artificial intelligence in use that have direct implications for higher education.
Applying Machine Learning to Scale Up Microcredentials
educau.se/dbadgeai
Penn State University Libraries married the areas of information literacy and competency-based education to create information literacy digital badges. They are piloting artificial intelligence to evaluate student work submitted for the badge, which provides real-time feedback for student responses.
CSUN AI Innovation Collection
educau.se/aiexp
California State University, Northridge held a yearlong faculty exploration program to explore AI and held a student competition to find new and interesting applications for AI. They also created an AI-powered chatbot, with the goal of helping students, faculty, and staff get 24/7 help to the most common questions anytime, anywhere.
Developing Virtual Patients for Medical Education
educau.se/vrmed
Virtual patients are avatar representations of human standardized patients controlled by AI so students can carry on a conversation using natural language. The system, from The Ohio State University, provides immediate feedback on student performance, allowing students to rehearse professional behaviors and interviewing skills prior to working with real patients.
For Further Reading
The following articles and resources are recommended for those who wish to learn more about artificial intelligence.
7 Roles for Artificial Intelligence in Education
educau.se/roleaied(Matthew Lynch, Tech Edvocate, May 5, 2018) This article outlines a variety of ways in which AI continues to be integrated into teaching and learning practices to increase student success.
Artificial intelligence (AI) Makes Learner-Centered Learning Successful
educau.se/aisucc(Open Access Government, June 1, 2018) Two professors from Chemnitz University of Technology outline how AI-enabled learning solutions can provide learner-centered education to students through real-time assessments.
Next Gen Robotics, Artificial Intelligence, and Education Informatization: The Future Is at TechCrunch Hangzhoueducau.se/robai(Technode, June 26, 2018) Five technology leaders in China discuss their experiences with AI, along with the trends moving AI forward and the ways in which the education sector can integrate it into existing online learning.