Ahead of the submission of their doctoral dissertation, a friend asked me to participate in a focus group to discuss digital intelligence and artificial intelligence. I prepared my answers in advance.
Focus Group Questions
The focus group questions are:
Grand tour question:
1. Please introduce yourself to the group and describe your educational background and career in your current position. CRQ
Most of my work has been as a technical writer for information systems serving government agencies and private enterprises. It happens that my degrees are in criminology and social science but I edit for the American Astronomical Society’s Historical Astronomy Division. Back in the 20th century, for five years, I taught technical writing at my local community college. [Added here: I completed a certificate in introductory use of ChatGPT via LinkedIn Learning as an assignment from my current employer, the University of Texas.]
Questions about Teaching and Education:
2. How do you perceive the role of data intelligence and AI education in preparing students for the future? CRQ
I understand education in data intelligence and artificial intelligence as aspects of good citizenship, along with literacy in mathematics and general science, as well as geography and politics, art and music, etc.
Gathering data from a wide range of sources and then evaluating each pool against a criterion places a special burden on the active citizen that exceeds the similar moral mandates for good citizenship from previous generations.
Before the 21st century, even in urban areas with more than one newspaper, most households chose just one based on political preferences. In that sense, there existed two assumptions: that the news was factual; its interpretation meshed with the values of the household. Given that most students K-12 follow the culture of the home, the descriptions of the world which were accepted by students aligned with that model. However, it is traditional in America that young people question those values when they matriculate to university education. While the specifics might change from nominally “conservative” to ostensibly “liberal” the mode of opinion formation was the same: the student found their facts and the meanings of those facts from established sources. All of that has changed.
Now, students must assume a special burden to first seek out valid (or least validatable) facts and then to give (or find) meaning in those empirical claims. Deeper still, the student must assume responsibility for the choices of original sources and tools of evaluation. In the previous generation, very few people used statistical methods (mean-mode-median; standard deviation; Xi-squared; p-values) to test the data they read about an effectively endless list of current events topics: inflation, unemployment, imports and exports, medical therapies, endangered species, etc.
Gathering basic information of events and dates is one task. Evaluating the economic, ecological, and moral contexts of those significant items is far more complicated and impactful. Whatever the worldview of the student among their family of origin, a large flow of unanticipated data and interpretation can have a life-altering impact.
Everyone has heard about ChatGPT. There are others from Microsoft (GPT-3), Google, (Bard LaMDA), Facebook (RoBERTa), and IBM (Watson), and many more competitors. How do you evaluate an AI? The choice must be made and becomes the responsibility of the student. First, however, the burden lies with the educator.
· “But with the sheer volume of data being collected, it became necessary to attach a value rating to the data itself, which led to a forensic approach to qualifying data assets by asking where they came from, when were they collected, and why were they collected in the first place.”
· “… data intelligence is specifically the collection of disparate pieces of data and using AI to determine what happened in the past and why, whereas data analytics is the use of that information to create actionable predictions of what may happen in the future.” --
· “… data intelligence first emerged as a means of gathering accurate background content for the purpose of more accurate and granular reporting. But with the sheer volume of data being collected, it became necessary to attach a value rating to the data itself, which led to a forensic approach to qualifying data assets by asking where they came from, when were they collected, and why were they collected in the first place. -- Hewlett Packard Enterprise. https://www.hpe.com/us/en/what-is/data-intelligence.html
3. How do you envision the future of digital literacy education in the ever-evolving landscape of AI and technology? SQ1
Educators at all grades and levels in every area are going to adopt AI tools the same as we gave up blackboards and wooden pointers to accept digital projectors and laser pointers—and, of course, computers. The only way to do that is to actively use an array of competing products to discover which are better or not-so-good depending on the needs of class, classroom, goals, and metrics. Microsoft Office and Adobe Suite have become common applications. However, we all learn to use an array of cooperation and team management tools, such as Teams, Slack, Confluence, Jira, Wrike, etc.
(In 2010, I had a professor whose handwriting was sometimes a challenge to decipher on the overhead. Caught short by the AV department, he had to write on the blackboard and his lettering was clear, distinct, artful. I asked him why he did not use the blackboard all the time. He showed us the dust on his hands and said that he was happy to be free of it.)
At an astronomical conference in 2021, one professor told us, “If they are not programming in Python, you are not teaching them astronomy.” That statement would have been unintelligible 100 years ago. Fifty years ago, it would have been rejected on the assumption that while computers may be useful for number crunching, serious astronomy is always [an] active engagement with a spectroscope or radio telescope. Now, we have so much data that we know only the proverbial tip of the iceberg. Professionals seek out and train adept amateurs, even people with no understanding of classical astronomy, who can mine the data.
Fifty years ago, few police officers had any education beyond high school. Today, 50% have associate’s degrees and 30% have bachelor’s degrees. (Data from 2017. Article from 2010: https://thehill.com/opinion/criminal-justice/504075-college-for-cops-studies-show-it-helps-their-behavior-stress-levels/ ) Those college programs include requirements in geographic information systems or symbolic logic or other classes previously considered far removed from walking a beat or driving a district. The new trends in data literacy and engagement with AI are just as strong and will impact just as many paid employments and lifestyle choices.
We know that you can lie with statistics, lie with numbers. The recent case of Prof. Francesca Gino is telling because it was a student who finally identified the fact that the data presented did not support the assertions in the narrative. These were peer-reviewed journal articles from a scientist who was considered a leader in the field of social science research. It means that everyone is responsible and capable of data literacy.
Questions about Digital Literacy:
4. How would you define digital literacy in the context of today’s AI-driven world? SQ1
See the quotes from Hewlett Packard Enterprise above (https://www.hpe.com/us/en/what-is/data-intelligence.html).Basic digital literacy is necessary to understand the past. With AI, digital literacy is the means of predicting futures.
Those are almost always statistical predictions: which are the likely outcomes; which ones are improbable? Digital literacy with AI requires knowing and understanding the tool sets as clearly as we know how to read the traffic signs when driving in a different city. AI does not yet have that common language.
That said, however, neither is this all as new as tomorrow. Right now, I am working on a short biography of Carl Sagan. I found an interview with NPR’s Talk of the Nation: Science Friday with Ira Flatow. Sagan explained the value in robots for exploring planets by pointing the technology of virtual reality. That was 1994, thirty years ago. The first chatbot was Joseph Weizenbaum’s Eliza program, which he warned against in his collection, Computer Power and Human Reason: From Judgment to Calculation (W. H. Freeman; 1967). So, we have 60 years, a lifetime or two generations, of this incremental development. Considering the evolution of printing, steam power, electrical power, aircraft, automobiles, or computers themselves, it is easy to see that this springtime of roots and shoots is going to become a full autumn harvest—whether we are ready for it or not.
Digital literacy will require knowing how to “read the signs” of products and processes, and navigate the landscapes of data without getting lost.
5. What are the key digital literacy skills and competencies that you believe are essential for students in the age of AI? SQ1
See the quote above: “If they are not programming in Python, you are not teaching them astronomy.” We all use computers and we do [not] think of navigating through Instagram, Reddit, or Facebook as programming but it is. Those other modes—Python, Java, etc.—are the digital literacy skills of the decision makers. Similarly, millions of people pour out billions of words via social media without knowing how to write a short story, a sonnet, or a limerick. It is an easy claim at that some level here and now, we are all digitally literate even though we do not have the same levels of competency.
The key skills and competencies for students in the age of AI remain what they have always been:
- Critical thinking
- A concerned and committed engagement with empirical evidence and logically consistent theory.
- Questioning
- Knowing when to accept the limits
Questions about Ethics and Ethical Considerations:
6. In your opinion, what ethical considerations should educators consider when teaching digital literacy in an AI world? SQ2
Technologies come and go. Ethics can change. Morality is constant. The invention of the cotton gin made slavery profitable and it was unethical to mistreat a slave but neither of those addressed the deeper problem. In order to make the best decisions about how to use AI tools, pupils and students must learn to ask the most important questions of themselves.
Questions about Student Engagement and Success:
7. How do you assess or measure students’ digital literacy skills, especially as they relate to AI? SQ3
You cannot write a meaningful examination for a subject unless you know the substantive material. With AI it might be possible to write an examination for any subject, accepting the warnings about the two basic limitations of AI: plagiarism [and] hallucination. So, any educator seeking to measure the work of their students must possess a deep understanding of the technology and its engagement.
8. In your experience, what are the main misconceptions or myths about AI that students often have, and how do you address them? SQ3
- The AI must be right.
- Hallucinations are easy to identify.
- Copying is permitted.
- We learn to trust authorities but we do not learn how to validate them. At some level, you begin pitting authorities against each other with no over-arching standards for judgment. How do you validate an AI. It is easy to find stories on the Internet about AI hallucinations and other failures but it is harder to find anything equivalent to the scientific method for substantiating the reliability of any AI product.
- AI hallucinations are claims that are obviously false, false by inspection. An article was never published; a person does not exist; an event did not occur. But those stories are easy to publish when everyone agrees on common facts. Students, in particular, are at an extreme disadvantage because they lack the life experience of academic learning. They have little choice but to believe what they are taught. The alternative is extreme skepticism, which is ultimately fruitless.
- One technique for learning how to write for publication is to get a journal notebook. Start with the works you like and copy passages by hand. You will work beyond this to find your own style. (Or so it is claimed.) The fact is that writers begin as readers. Educators properly require that their students read and then rephrase what they have read in their own words but the standards for that are not objective. “President Wilson was adamant that the United States must support the United Kingdom in its war against Germany.” Does changing the word “adamant” to “insistent” prevent a charge of plagiarism?
PREVIOUSLY ON NECESSARY FACTS
John Kemeny Knew: We Shall Have Computed
ArmadilloCon versus Artificial Intelligence