The Tapestry of Technology
A.I. will disrupt work as we know it. Ursula Franklin's 1989 lectures help us to understand how and why.
If you’re anything like me, you’re simultaneously enthralled and exhausted trying to keep up with what seems to be an ever-accelerating pace of development in artificial intelligence right now. Only four months on the heels of the initial release of ChatGPT, OpenAI released their updated GPT-4 model a few weeks ago. GPT-4 improves upon the already impressive abilities of the original GPT-3 model which powered ChatGPT at its release and now routinely performs at a high level on a wide range of professional exams. And hot on the heels of GPT-4, last Thursday OpenAI announced the rollout of plugins, enabling ChatGPT to interface with external tools like Wolfram|Alpha and browse the web.
While on the one hand, I feel overwhelmed trying to keep up, I still can’t stop talking about A.I. As I continue to use these tools, I am increasingly convinced that they will drastically reshape the world as we know it, especially in the domain of knowledge work. As an educator, I feel a responsibility to stay on or near the cutting edge to partner with my students and colleagues as we navigate this new world together and find our way through the various levels of disruption that are here and will continue to come.
How do we continue to move forward? By looking back. This week I want to present two frameworks that have helped me to see the issues more clearly: understanding technology as a tapestry and identifying the difference between holistic and prescriptive technologies.
Sitting with questions is more important than ever
One of the things I appreciate most about the prototyping mindset is that it gives space and permission to sit with questions without needing to jump to solutions. Answers are brittle. They often depend on specific conditions or environments to be valid.
Good questions, on the other hand, are robust. While they can’t be the end of the process, any designer or engineer knows that getting the question right is the key to developing a good answer. And as I often remind myself, the right answer to the wrong question is the wrong answer.
So what are the questions that we should be asking about A.I.? The scope of the related issues is wide, but I think it is valuable to take a step back and try to get a fifty-thousand-foot view of the situation.
As I see it, there are several categories to consider as we move forward. Of course, the application of these ideas will vary strongly depending on the specific application and its context, but broad categories are helpful nonetheless. A few of the questions that I have been pondering lately include:
Regulation: Should we regulate the development of A.I. tools? If so, how and who?
Individual Impact: What is the impact of these technologies on us as persons? How are these tools shaping us and promoting or preventing our flourishing?
Societal Impact: What are the impacts of A.I. tools on communities of people?
Invisible Bias: How might the selection of data and the human reinforcement used to train these models emphasize certain underlying values or perspectives over others?
Impact on Work: How are these tools likely to disrupt work and specific sectors? How should we respond?
Unintended Consequences: What are the potential unintended consequences, both positive and negative, that might accompany the continued development and deployment of A.I. tools?
Boundaries of Use: What things could A.I. do that it should not do? Who should decide and how?
Intellectual Property & Data Ethics: How do we navigate the issues surrounding A.I. and its training data? How is the data we are providing these tools being used?
There is nothing new under the sun
As I’ve been thinking more about these questions, I’ve been curious to explore and think more about how others throughout history have thought about similar eras of technological disruption. While the specific contours of our current conversations about A.I. may be unique, there are many aspects of its potential impacts on society that can be modeled from the past.
In this vein, I am learning a great deal from Ursula Franklin’s 1989 CBC Massey Lectures1 which she later published in a book titled The Real World of Technology. In her set of lectures, she gives many helpful examples of how past technologies have developed and provides what I think is a helpful framework for considering the impact of A.I.
From her lectures, two ideas have been resonating with me this week.
The Tapestry of Technology: Technology must be understood as part of a tapestry, integrally connected to other elements of society. When one part of the tapestry is altered, the entire woven structure is impacted.
The Shift from Holistic to Prescriptive Technologies: The transformation of holistic technologies that integrate development from beginning to end into prescriptive technologies that efficiently slice work into small steps comes at a substantial cost.
1. The tapestry of technology’s impact
One of my biggest takeaways from Franklin’s lectures is the way she describes the impact of technology as a tapestry or web of interactions. While some think about technology as a complex system with many different actions and reactions, the imagery of a tapestry draws specific attention to the idea of patterns or designs that run throughout a system connecting various elements within it and the way the strength of the overall tapestry relies upon the regularity of its pieces.
In the context of technology, the patterns in the tapestry span many different areas and interests. Inventors and creators are one part. They in turn connect to the individuals who use their inventions and the communities composed of those individuals. This extends even to those who never directly use the technology themselves but are shaped by it. Think of the ways that the automobile-centric design of Los Angeles impacts even those who do not own or use a car. Or the way that social media shapes the world even for those who do not use it. Because of these deep connections, the impact of technology ripples through society and influences even those who are not directly creating or using the technology itself.
Franklin writes:
The real world of technology is a very complex system. And nothing in my survey or its highlights should be interpreted as technological determinism or as a belief in the autonomy of technology per se. What needs to be emphasized is that technologies are developed and used within a particular social, economic, and political context. They arise out of a social structure, they are grafted on to it, and they may reinforce it or destroy it, often in ways that are neither foreseen nor foreseeable. In this complex world neither the option that “everything is possible” nor the option that “everything is preordained” exists.
This framing helps to highlight the ways that technology and culture are woven together. As such, the values and inclinations embedded within technologies have a widespread impact, whether they are explicitly promoted by the inventors or implicitly leveraged by the societal and economic incentive structures surrounding them.
We must consider the tapestry of technology as we ponder the current development of A.I. and its potential impacts. What are the core values driving its development? Who will benefit? Who will be adversely impacted?
While we may optimistically hope that the widespread adoption of A.I. will create opportunities for growth and human flourishing, it is assuredly the case that some work will be made obsolete. Moving forward without pondering both the clear and hidden connections is reckless at best.
2. Holistic vs. prescriptive technologies
One of the beautiful ideas that Franklin threads throughout her lectures is the idea of holistic vs. prescriptive technologies. This distinction draws out how work is being done instead of focusing only on what is being done. I think this lens is particularly enlightening as we consider how to integrate A.I. into our work.
Holistic technologies are normally associated with craftsmanship, where a single artisan like a cook, painter, or potter controls the process from beginning to end. While this does not mean that people do not work together in the process, it does mean that the work weaves together in such a way that it leaves the individual worker in control of their process of creation or work.
Prescriptive technologies on the other hand are fragmented. No one person controls the production from beginning to end. Instead, the technological process is broken down into steps and individuals have control over only their small part of the process.
Perhaps the classic example of prescriptive technologies is the assembly line. The product being built moves from one station to another as various parts are installed or processes take place. However, prescriptive technologies extend beyond processes of material production or assembly. For instance, consider how many management and administrative processes model themselves in a similar way.
Franklin points out that these prescriptive technologies are often touted for their effectiveness and efficiency, but this comes at a cost, what she terms an “enormous social mortgage.” By leaning into these prescriptive frameworks, we foster a culture of compliance that promotes a single right way of accomplishing a task. Unfortunately, while that one way might be the best on the merits of efficiency, it likely comes at a cost to human flourishing and well-being. In this frame, the goal is focused on material productivity as opposed to fruitfulness.
I am of two minds when it comes to thinking about A.I. and holistic vs. prescriptive technologies. In one sense, A.I. can continue to advance in a prescriptive framing, further fragmenting our work into ever smaller pieces that we will farm out to automated systems.
On the other hand, I wonder if the suite of A.I. tools that are beginning to develop (e.g., ChatGPT + Midjourney) might serve to increase our ability to work holistically with A.I. as an assistant, extending our capabilities for holistic work. For example, singlehandedly creating a smartphone app is a challenging task for any one person. There is the coding required to power its operation, the graphic design needed to polish the experience, and the copywriting needed to promote and advertise the app. However, as Ethan Mollick demonstrated in one of his recent posts, the current stack of A.I. tools enables a single person to control the process of creation from start to finish.
So what?
There are many questions facing us as A.I. continues to rapidly improve. Some think that the specter of A.I. is cause for grave concern, warning that the trajectory we are on is likely to destroy the foundations of our society. Others are more sanguine, emphasizing that A.I. is a tool and that arguing that A.I. will lead to species extinction is premature without presenting a pathway.
Whichever way you tend to lean, it is critical to remain curious and questioning. As I continue to ponder these issues, I think that Ursula Franklin’s imagery of a tapestry and the contrast between holistic and prescriptive technologies helps us see the issue from a variety of angles.
Please join me in sitting with these complicated questions and sharing your thoughts, responses, or curiosities by commenting or sending me an email. Thanks as always for reading.
The Book Nook
I would strongly encourage you to engage with Ursula Franklin’s lectures either by picking up a print or digital copy or listening to the original recordings. They are relatively digestible in small chunks and filled with thought-provoking ideas.
A few quotes
Technology is more than a tool.
Technology is not the sum of the artifacts, of the wheels and gears, of the rails and electronic transmitters Technology is a system. It entails far more than its individual material components. Technology involves organization, procedures, symbols, new words, equations, and, most of all, a mindset.
Prescriptive technologies lead to the narrow idea of a “right way.”
Today’s real world of technology is characterized by the dominance of prescriptive technologies. Prescriptive technologies are not restricted to materials production. They are used in administrative and economic activities and in many aspects of governance, and on them rests the real world of technology in which we live. While we should not forget that these prescriptive technologies are often exceedingly effective and efficient, they come with an enormous social mortgage. The mortgage means that we live in a culture of compliance, that we are ever more conditioned to accept orthodoxy as normal, and to accept that there is only one way of doing “it.“
How will we grapple with the way technology separates us?
One illustration of technologically induced human isolation: When I go to work in the morning I often meet a neighbour and her ten-year-old daughter. Every day they walk side by side to the bus stop, each plugged into her own Walkman, isolated from one another and from the rest of the world. Such is the real world of technology. The question that lingers on in my mind is this: How will our society cope with its problems when more and more people live in technologically induced human isolation?
Tools redefine problems.
Tools often redefine a problem. Think, for instance, of speeding and radar traps. Let’s go back to the purpose of speed limits. They were instituted to enhance safety, not to produce criminality. One way of enforcing speed limits used to be the judicious presence of clearly marked police cruisers on our highways. The police drove at the speed limit and by this tactic brought the traffic pattern into compliance with the regulations. The tool of radar traps brought another dimension into the situation. The emphasis shifted from common safety to individual “deterrence.” It was felt that the fear of being caught and fined would be a better way of enforcing the regulations. Next came a technological option of avoiding the radar trap, using what’s commonly called a “fuzz-buster.” Now the motorist, concerned less with safety than with criminality, buys an avoidance device, whether it is outlawed or not. The next player in the speeding game is a device for law-enforcement officers to detect the presence of a fuzz-buster. And now there seems to be a new generation of widget on the horizon which those with a fuzz-buster can use to detect the counter-technology of law enforcement. And so it goes.
The Professor Is In
I am excited to be giving a talk for the Harvey Mudd community in a few weeks titled “The Power of the Prototyping Mindset: How Embodied Questions Can Help You Build a Meaningful Life and Career.” In the talk I will unpack some of the underlying philosophies of the prototyping mindset and offer some thoughts on how I’ve used these ideas to make progress in my life and how you might consider implementing them to make progress in yours.
If you are interested, you can find more information at the webpage here. A recording of the talk should be posted soon after it takes place and I’ll be sure to share it in a future edition of the newsletter.
Leisure Line
Back at it this week with some New Haven-style dough for Friday night pizza night. Had to pull out my favorite potato rosemary pie ala Sally’s Apizza.
Still Life
Gotta wonder what they were thinking when they put this cover on… I guess the CAT5 cables will still fit through somehow?
It is a real treat that in addition to reading these lectures, the original recordings are available online for free: https://www.cbc.ca/radio/ideas/the-1989-cbc-massey-lectures-the-real-world-of-technology-1.2946845