A Resource for Artificial Intelligence

information technology verbiage
Prev Next

Welcome to the College of Staten Island’s website on Artificial Intelligence: AI@CSI. This website can be used as a resource for faculty, staff, and students who would like to be more informed about AI. Are you exploring AI in teaching and learning? What about your research? Are you concerned about ethical practices and associated challenges?  Are you knowledgeable about how students use AI? Are there benefits for administration? What kind of AI tools are available? There is so much to cover as the world of AI continues to evolve.  AI@CSI is only the beginning of your adventure!

 

What is AI?

Artificial intelligence (AI) refers to the ability of computers to perform tasks that typically require human intelligence such as reasoning, problem-solving, and decision-making. AI does this by using algorithms that compute and process large volumes of data and extract patterns. AI then makes predictions or decisions based on these patterns. 

Generative artificial intelligence is a specific subset of AI, which focuses on creating content such as text, images, video, music, etc. These are created in response to what the user requests. Generative AI models are designed to learn the patterns and structure of their input training data and generate new data with similar characteristics.  

Generative AI tools can easily generate a wide variety of human-like outputs; therefore, they have the potential to radically transform the way we approach content creation.  It is important to be very cautious since AI outputs are derived from undocumented data sources, infringe on intellectual property, and are prone to error.  

* Rutgers University

Best Practices

How Do Students Use AI Faculty Use of AI for Teaching Research Administration Under the Hood Universal Design

Use of AI for students can be advantageous to the learning experience. Some examples: 

AI and Assessment: Where We Are Now | AACSB 
Temple, students use free ChatGPT accounts when they brainstorm and write rough drafts. As part of their assignments, students must submit transcripts of their conversations with ChatGPT as well as reflections on how the process worked and whether the tool was helpful. 

Also at Temple, students use ChatGPT to summarize a case. Then they add line-by-line comments about where they agree and disagree with the chatbot, before making a final analysis. As an additional step, student peers review each other’s work. 

Links:

AI can be beneficial in helping instructors create rubrics and improve the quality of their course syllabus and assignments. Consider leveraging AI as a toolbox (Edge.net 2024). Teachers are experts, but  AI can assist by finding new ways to improve the learning experience. “AI is not a search engine; it is more like a knowledgeable colleague, it is more about prompt engineering and having a conversation that fine tunes the results. Faculty should see AI as an idea generator that could be leveraged and helpful with many aspects of the classroom and beyond (Watson, 2024).

Some of the challenges associated with AI relate to:

  • Resistance to change: where faculty and staff may resist adopting AI due to unfamiliarity or concerns about job roles.
  • Effectiveness Evaluation: Measuring the impact of AI on student learning outcomes and overall educational quality is essential. Developing appropriate evaluation metrics can be challenging
  • Pedagogical Integration: Integrating AI seamlessly into existing teaching methods and curricula requires thoughtful planning. 

Hanover Research presents strategies for managing Artificial Intelligence (AI) in higher education settings. This report discusses the unique challenges that AI presents as well as guidance for management and faculty to effectively engage with artificial intelligence.

How is Research Conducted in an Academic Setting?  
Much of the research and scholarship conducted in an academic setting follows a time-honored approach known as The Scientific Method. This approach allows a researcher to try to answer research questions using a prescribed set of steps where answers to research questions are derived from observation and gathering existing information, developing an explanatory hypothesis, gathering data to test the hypothesis, analyzing the data to draw conclusions, and sharing the approach and results through publication so that the hypothesis can be retested by others.  Often, it may take researchers years to go from their initial question formation to the publication of data that addresses the hypothesis.

How does Artificial Intelligence Impact Research? 
Artificial Intelligence can impact each step of the Scientific Method and widespread access to AI has the potential to speed up the research process and allow researchers to reach conclusions more efficiently. As a researcher develops a question, AI can help to direct the researcher to seminal works in the field and existing theories around the question at hand.  

Artificial Intelligence using Large Language Models Can Aid Research 
Large Language Models like Chat GPT and Bard have been trained using libraries of scientific works and can help researchers to understand the appropriate terminology and current state of understanding in a field.  Of course, researchers should be expected to read the appropriate scientific publications in detail before drawing conclusions about the work, but AI tools can help to get researchers started in their investigation in the way a review article might.  

Using AI to Refine the Hypothesis
AI can help researchers to refine hypotheses by providing feedback about the feasibility of planned experiments and providing insight about the number of datapoint needed and how to collect them efficiently.  Researchers tend to be knowledgeable about their own fields, but there may be techniques and approaches in other fields that researchers can benefit from.  Large Language Models may help researchers learn from these tangential fields. 

AI and Data Collection
AI tools have been helping researchers through a variety of programs for at least the last decade.  For example, AI software can help monitor environmental conditions to help a weather balloon navigate to an area of interest.  In addition, computer vision-based programs can use machine learning to detect and measure features of interest in images acquired from satellites or microscopes.  One further example leverages AI based tools detect and transcribe human speech and analyze speech and text for sentiment.   Perhaps the largest research advancement from AI comes in the form of data analysis where tools can help to identify patterns in large datasets and apply advanced statistical techniques and visualizations of data.  Researchers can learn from AI and assist with their findings.

AI productivity

AI can assist with streamlining tasks and increase efficiency. It is a powerful tool, when used correctly will aid in handling daily tasks.   Remember, it’s AI so be sure to use this as a guide and use the generated content as the final output.  Let’s look at a few examples: 

  • We all know about Virtual Assistants such as Alexa, Siri, or Microsoft Cortana, or Samsung Bixby. Have you ever considered using these VA’s to handle tasks such as managing calendars and sending reminders?  
  • Simple AI scheduling tools can be used to automatically detect availability of meeting participants 
  • How about using Email management tools by sending automated reminders, filtering, and sorting 
  • Transcription tools that can automatically transcribe audio and video files. 
  • Take advantage of ChatGPT! Use AI to answer questions, generate email content when you are drawing a blank slate, and maybe even condenses complex data into easy-to-understand summaries. 
  • Customer Support through chatbots 
  • Generating reports 
  • Provides for better collaboration and communication 

And let’s thank Bing for the following resource:   Can artificial intelligence actually increase human productivity? 

  • How AI Transforms the Workplace  
    Discover the profound impact of AI on productivity, with insights into how it’s revolutionizing industries and reshaping the future of work. 
  • AI and Automation: The Future of Work 
    Explore the opportunities and challenges presented by AI and automation. 
  • Maximizing Efficiency with AI 
    Dive into the specific ways AI applications drive efficiency for organizations.
  • AI in Action: Real-World Success Stories  
    Read about companies that have successfully integrated AI into their workplaces. 
  • Preparing for an AI-Driven Workplace 
    Equip yourself with the knowledge to navigate the AI-enhanced workplace.
  • AI Tools and Technologies 
    Get acquainted with the latest AI tools and technologies that are guiding, organizing, and automating work in next-generation workplaces.
  • Budget Constraints 
    Implementing AI solutions often involves costs related to software, hardware, training, and maintenance. 

Additional Resources 

Exploring the Depths of Artificial Intelligence: Understanding Functionality, Learning Processes, and Ethical Implications 

Artificial Intelligence (AI) stands at the forefront of technological innovation, reshaping industries, challenging human capabilities, and raising profound ethical questions. In this section of our website, we embark on a comprehensive exploration of AI, aiming to elucidate its functionality, learning processes, and ethical implications. Through an interdisciplinary lens, we synthesize technical insights with ethical considerations, shedding light on the evolving landscape of AI and its implications for society. 

Blurred Lines

AI Network

Artificial Intelligence has emerged as a transformative force in modern society, permeating various sectors and revolutionizing the way we live and work. Understanding the intricacies of AI's functionality, learning processes, and ethical implications is paramount as we navigate the increasingly complex interplay between technology and humanity. 

The term artificial intelligence prompts us to reconsider the foundations of consciousness, drawing parallels between the neural network architecture of AI and the biological complexity of the human brain. While humans have long asserted their consciousness based on subjective experiences, the emergence of AI introduces a paradigm shift, blurring the lines between organic and artificial intelligence.

 

Mirroring the human brain 

AI Layer of Nodes

At the heart of AI lies the neural network, a structure that mirrors the complexity of the human brain. A neural network is designed based on the human brain, except that what we call nodes and linkages, are called networks of neurons and synapses in the human brain. Comprising layers of interconnected nodes, neural networks are adept at processing vast amounts of data and identifying intricate patterns. Through simplified examples like image recognition, we gain insight into the inner workings of neural networks and their role in AI's functionality. 

 

How does AI Function 

Ai a dog

One of the most common questions we encounter is how AI functions. Essentially, AI learning processes are defined by iterative training techniques that leverage large datasets.

Supervised learning, a fundamental method in AI training, involves feeding labeled data to the neural network, allowing it to learn and enhance its predictive accuracy.

The gradient descent algorithm aids in adjusting neural network parameters to minimize prediction errors, thereby improving performance. 

In the above example, when an AI is given an image of a dog, the image is decomposed into data pieces that pass through each node. The data progresses from the first layer of nodes to the second, and through successive layers until it reaches the final layer. The neural network then computes the value of each final node, and based on these values, it determines that the image depicts a dog.

AI Generated Content – Copyright Infringement
AI's ability to replicate artistic styles and content has sparked debates surrounding creativity and plagiarism. While AI-generated content may bear resemblance to human-created works, it operates on principles of pattern recognition rather than originality. Legal disputes regarding AI-generated content underscore the need for nuanced discussions on copyright law and artistic expression in the digital age. 

Scientific Calculations
AI's potential to solve complex mathematical problems through pattern recognition is a topic of intrigue and speculation. By approximating patterns in data, AI can infer solutions to problems that defy conventional mathematical approaches. Case studies on AI's role in predicting protein folding demonstrate its efficacy in addressing scientific challenges and advancing research frontiers. 

Human Cognition
The comparison between AI and human capabilities evokes questions about the limits of technology and the essence of human intelligence. While AI excels at pattern recognition and data processing, it lacks the nuanced understanding and creativity inherent in human cognition. Ethical considerations surrounding AI's impact on the workforce and societal structures necessitate careful reflection on its implications for human society. 
The exploration of AI consciousness delves into the realm of science fiction and real-world analogies, challenging our understanding of intelligence and sentience. As AI evolves, questions of autonomy, accountability, and ethical oversight become increasingly pertinent. The intersection of technology and humanity compels us to grapple with profound questions about the nature of consciousness and the ethical implications of AI's autonomy. 

Looking Ahead
Future research in AI will undoubtedly explore new frontiers in neural network architecture, ethical frameworks, and societal implications. As AI continues to evolve, interdisciplinary collaboration and ethical considerations will be instrumental in shaping its trajectory and maximizing its potential for the betterment of society. 

Bibliography

Links

There are several considerations regarding the use of artificial intelligence (AI) in the classroom context and within the standard of universal design for learning.  
The most crucial point to start: The underlying features of AI platforms must be accessible to individuals with a disability! 

Where WCAG comes in: 

Web Content Accessibility Guidelines (WCAG) are standards created by the World Wide Web Consortium (W3C) as a special initiative. Though not related to the Americans with Disabilities Act (ADA), both public and private entities have adapted it as the technical standard to adhere to accessibility requirements set forth by the ADA. 

When understanding how WCAG is successfully implemented, it is helpful to think of the acronym POUR – Perceivable, Operable, Understandable, and Robust. Any AI software must be: 

  • Perceivable: Users must be able to perceive, or receive, the information in some way – meaning, it cannot be invisible to all the senses 
    • For example, low-vision/Blind individuals must have access to alt-text for any images that AI generates 
  • Operable: Users must be able to operate the software within their capacities (the software cannot require an action that the user cannot perform) 
    • For example, must account for individuals who cannot use traditional mouse / keyboard 
  • Understandable: Users must be able to understand the information and operational tasks being presented to them 
    • For example, appropriately targeted language / reading-level 
  • Robust: Content must be robust enough that it can be read and interpreted reliably by all user agents, including assistive technology tools – and be able to adapt to the continual changes in technology 
    • For example, support some outdated browsers/systems but be consistent with the latest standards 

It is important that any AI software be accessible and user-friendly to people with various needs, such as visual difficulties, hearing difficulties, and learning difficulties. WCAG compliance ensures that the applications fit these needs. 

If you would like more in-depth information regarding WCAG standards, please refer to the two links below: 

Beyond the fundamentals of making sure that the AI user interface is accessible to individuals with disabilities, it is important that AI be used mindfully and meaningfully to enhance the learning experience. AI is not a substitute for learning the basic skills of a topic or objective.  A template has been developed by AI for Education and the North Carolina Department for Public Instruction, known as EVERY, to guide students in using AI to supplement learning: 

  • Evaluate the initial output to see if it meets the intended purpose of your needs 
  • Verify facts, quotes, figures and data using reliable sources to ensure there are no hallucinations or bias 
  • Edit your prompt and ask follow up questions to the AI to improve output  
  • Revise the results to reflect your unique, style, and tone. AI should not be the voice writing the information 
  • You are responsible for everything you create with AI. Be transparent in your work and citations about how you’ve used the tools 

Please visit AI for Education for more information about the EVERY acronym and its use and guidance in the classroom. 

And finally, AI doesn’t have to just be helpful for a singular lesson...it can be helpful to enhance the entire classroom and learning experience!  

Cornell University lists a number of ways that AI can be used for more flexible online assignments to enhance learning outcomes for diverse learners: 

  • Accommodates different formats for various assistive technology software 
  • Presents with fewer distractions than the physical classroom 
  • Provides extra time to read and remember/process tasks 
  • Allows for engagement with curriculum when student cannot physically be present 
  • Are ENL learners who may need additional time to read instructions 

Links: 

Ultimately, AI in the context of UDL means that AI must enhance the learning experience – both in terms of design accessibility for different learning styles, as well as a tool to enhance one’s learning, rather than a substitution. It is crucial that through instruction, students learn both the benefits and drawbacks of AI to best utilize it within the classroom.

Considerations and Resources

Ethical Considerations AI Tools Available/ ChatGPT Data Privacy and Sharing Who is Using AI @ CSI? Resources What Does the Future Look Like?

Bias  
Bias is a key ethical pitfall for large language models (LLMs), which can absorb prejudices from their training data to perpetuate harmful stereotypes against traditionally marginalized communities. In educational contexts, without responsible development focused on debiasing, AI tutoring systems or admissions screening tools can easily discriminate against certain student demographics based on attributes like race, gender or socioeconomic background. 

Disruptive Impact 
The disruptive impact of generative AI across domains like writing, analysis, and creative expression could fundamentally reshape pedagogical approaches and curricula. As AI automates aspects of these disciplines, education may need to pivot towards cultivating uniquely human skills like creativity, critical thinking, and complex reasoning to prepare students for the AI-augmented workforce. Simultaneously, AI literacy through ethics education will be vital to ensure the responsible development and use of these powerful technologies. 

Environmental Impacts
Environmental impacts from the immense energy demands, electronic waste, and resource usage in developing generative AI also cannot be ignored. Training large neural networks requires immense computational power. Some estimates indicate that training a single large AI model can have a carbon footprint equivalent to nearly five times the lifetime emissions of the average American car (including the making of the car itself). The manufacturing of hardware infrastructure (GPUs, servers, cooling systems etc.) to support AI training at scale relies on the mining and exploitation of natural resources, which can have damaging environmental impacts. Innovations in energy-efficient hardware, renewable energy, green AI practices, and sustainable data centers are needed to minimize the carbon footprint. 

Other Impacts
This is by no means a comprehensive list of ethical concerns in AI. Gaps in access to expensive high-quality tools, job redundancy among information professionals, as well as copyright and privacy concerns also loom large. 

Job Displacement: 

Unequal Access to Technology: 

Resources

AI Tools

There is a plethora of tools available for AI. The CSI Library created a comprehensive list that provides insight on Generative AI Core Competencies as well as Generative AI Tools for Educators. 

It's important to note that the field of AI in education is rapidly evolving, and new tools are constantly being developed. Educators should always evaluate these tools carefully for effectiveness, privacy concerns, and alignment with educational goals before implementing them in their classrooms.

ChatGPT

While rudimentary chatbots have existed for several decades, ChatGPT represents a step forward in artificial intelligence driven computing, allowing users to rate the responses it gives to prompts, which enables the algorithm that generates those prompts to be further refined with continued use.  

A main privacy and data security concern in generative AI systems stems from the large datasets used to train large language models (LLMs), which may contain personal or sensitive information scraped from the internet without consent. Techniques like model inversion and membership inference attacks can be used by hackers to reconstruct or identify specific data points from the training data, exposing private user information. LLMs can also create highly realistic fake identities, images, audio, and more, which could be maliciously used for misinformation campaigns, impersonation, or fraud.  

Generative AI poses specific risks for American colleges and universities. Academic research data containing personal details could also be compromised. In efforts to combat cheating, students’ work may be entered into plagiarism-prevention systems without their knowledge or understanding, raising copyright concerns. Students and faculty alike may not know what they are agreeing to when they create accounts within these systems, allowing companies to use their sensitive queries or other system interactions to continue to train the software. 

Educational institutions must adopt a comprehensive strategy to mitigate the privacy and data security risks posed by generative AI. This should involve robust data governance policies, de-identification protocols, and ethical AI frameworks that enshrine principles like privacy and transparency. They need to restrict access to sensitive data, update academic integrity policies to prohibit misuse of generative AI for cheating, enhance cybersecurity defenses, and incorporate AI ethics and privacy education across curricula. Establishing cross-functional AI governance committees, collaborating with industry on best practices, and advocating for clearer regulations are also crucial. 

Faculty/Research 

Administration 

  • Office of Technology Services application developers.

More to come....

Students Use of AI

Faculty Use of AI in Teaching

Administration

Under the Hood

Universal Design

Ethical

Tools

Data Privacy

One thing is clear…

The future is AI! Continued economic interest has surrounded artificial intelligence. There has been a 14% increase in the amount of AI startups since 20001, and according to a survey that IBM conducted in 2023, 42% of businesses have integrated AI into their operations, with another 40% considering implementation2. Generative AI, in particular, is estimated to potentially add $4.4 billion to the economy3.

There are several industries that have seen the most significant impact from the emergence of AI1,2,3,4:

  • Education
    • AI can be integrated in several ways throughout the learning experience – running simulations, providing personalized learning experiences for various learners, encouraging technological training, and giving students an opportunity to develop strong ethical skills. This can be done at both the secondary and tertiary level.
  • Healthcare
    • Through data analysis capabilities, AI has begun to identify ailments and disease more accurately, streamline drug discovery through simulating clinical trials, and virtually monitor patients. It has the vast potential to quickly and efficiently analyze massive datasets from individuals’ physiology.
  • Finance
    • AI will power more and more “frictionless” currency, such as blockchain, cryptocurrency, and distributed finance – all forms of digital finance that are integral to today’s world. Additionally, AI is used to detect fraud, conduct financial audits, and evaluate potential customers for loans. Finally, its use is getting incorporated by investors and influencing their market decisions, due to its ability to analyze millions of data points.
  • Media
    • Generative AI intelligence has gained traction for developing algorithms that create novel content, including audio, code, images, text, simulations, and videos. AI has additionally been used
  • Customer Service
    • AI has the capacity to supplement customer service agents and glean key insights about customer satisfaction through its use as chatbot and virtual assistant.
  • Foreign Policy
    • AI will be integrated to make the United States more economically resilient and geopolitically strong and competitive in today’s global economy
  • Transportation
    • Self-driving cars are becoming increasingly commonplace… and AI has also begun to insert itself into the tourism industry, creating virtual guides for travelers on various locations.

There are also questions about when AI’s capabilities will match humanity’s – a term known as “General Artificial Intelligence”. Projections can vary widely for this, from several decades to hundreds of years6. There is no true consensus on the timeline, but what is understood is that these developments will occur and impact our society6.

Finally, as noted above, there also stand to be significant impacts on higher education through AI advances2,5:

  • AI creates opportunities for personalized learning and ways to suit various learning differences – including but not limited to Deaf/HOH, Blind/Visually Impaired, and students with learning disabilities
  • Provides automation of administrative tasks, such as prompting, feedback, grading, and assessment – leaving more time for educators to develop curriculum
  • Equips graduates with skills such as creativity, coding, and prompt engineering abilities
  • Gives admissions offices and administrators the ability to anticipate enrollment trends and optimize student recruitment efforts

How can we tackle integrating AI into the future?

  • Faculty training is crucial5! Creating seminars to introduce AI to emphasize benefits and educate on limitations, as well as demonstrate case studies on what to expect.
  • Preparing students to incorporate AI into their academic and professional work, as AI is estimated to be utilized to some degree in tasks in the majority of professions and careers5.

Sources:

1: In-Depth Guide to Future of AI in 2024, According to Top Experts (aimultiple.com)

2: The Future of AI: How AI Is Changing the World | Built In

3: What is the future of AI (Artificial Intelligence)? | McKinsey

4: The Future Of AI: 5 Things To Expect In The Next 10 Years (forbes.com)

5: Zouhaier Slimi: The Impact of Artificial Intelligence on Higher Education: An Empirical Study; European Journal of Educational Sciences, March 2023 edition Vol.10 No.1 ISSN: 1857- 6036

6:AI timelines: What do experts in artificial intelligence expect for the future? - Our World in Data

 

For more in-depth information, please take a look at the following pages: