The Tantalizing Promise of Generative AI: The Integration of Everything
The promise of new large-scale systems and multi-disciplinary innovations powered by Generative AI is founded on its potential for unprecedented knowledge integration. How would it positively impact our world?
Like many of us, I have been trying out chatGPT and its ilk; been listening to the implications of Generative AI (“GenAI”) from pundits — in startups to large corporations, academia, neuroscientists, and linguists.
While chatGPT’s failures in answering questions and hallucinations have been spectacular as documented, I have been impressed by the answers it does get right.
In those answers lie its potential — the promise of what it could be when GenAI or its successors get things right; the promise of AI-powered large scale integration of everything — of information, services, products, industries, and more.
But first, an excerpt from my q&a with chatGPT.
A Question and its Answer
I asked chatGPT to find the ties across three domains — neuroscience, behavioral economics, and public policy. The answer matched my own realization after 50-hours of coursework and research in these fields and more reflection afterwards!
The answer was quick. No second was wasted “thinking”.
The three fields are connected through their shared interest in understanding and influencing human behavior and decision-making processes.
I won’t go into the long technical discussions on how chatGPT arrived at that answer. I’ll focus on the implications of a technology that connects the dots across all human knowledge — especially domains we don’t think of as related in our normal discourse.
This is not our garden-variety data integration of the last few decades where we extracted, scrubbed, transformed, and unified data from various sources — largely homogeneous sources. Nor is it application or process integration. It is not the same-old workflow integration and automation.
It is knowledge integration across heterogeneous data types and sources (“multi-modal” data) enabled by large-language models (LLMs).
Let’s look at the implication of such large-scale knowledge integration in four dimensions:
- Integration of Technologies
- Integration of Disciplines (Fields of Study)
- Integration of Business Functions
- Integration of Industries
1. Integration of Technologies
If chatGPT could rapidly make connections between disparate disciplines as seen above, what could it do across the multitude of STEM (Science-Technology-Engineering-Math) fields?
Below is a simple illustration of how integration of knowledge across STEM fields unearthed technological innovation historically.
Emerging examples of this integration include blending biotech and material science research to arrive at new drugs and drug delivery, accelerating the discovery of new cures and precision healthcare by combining nanotech and biotech etc.
GenAI has the potential to give these innovations a fillip via easier access for more people the knowhow to integrate across disciplines. Democratization of knowledge integration, so to speak.
2. Integration of Disciplines (Fields of Study)
What if we extended that integration of knowledge beyond technologies to cover not just STEM disciplines, but humanities and sports too?
Would it not make answers to questions such as “why do women have more ACL injuries and what can be done about it” easily accessible for faster implementation of coaching programs?
Would it not analyze Bach’s music or Picasso’s paintings and unearth underlying patterns of human creativity (as described in Gödel, Escher, Bach by Douglas Hofstadter).
Would it not help us find naturally occurring patterns to design new products — such as the king fisher inspired bullet train.
History is full of nature-inspired or artist-inspired product designs. GenAI could catalyze new designs.
Beyond potential positive effects of GenAI-based integration across STEM, humanities, and sports, we are on the cusp of amplifying human creativity.
The debate on whether AI is creative and whether it is creating new knowledge is a different discussion. Even if GenAI is merely generating “stuff” by the recombination of existing knowledge or micro-curation of existing content, it still has enormous potential.
3. Integration of Business
New technologies always automate business processes and workflows. GenAI is no exception. We are already seeing the automation creep-up to even jobs seen as creative such as copy-writing.
However, the potential for GenAI to integrate business goes farther than prior technologies. Imagine having a GenAI copilot for your entire marketing function or sales function? Could it break the siloed sub-functions that exist within them?
Going further, could it integrate across business functions? E.g. Demand generation in marketing tied to sales and services. These are already being promised and implemented to some extent.
Could we extrapolate this further and realize an enterprise-level copilot for the C-suite to manage the business? E.g., can we forecast a supply-side risk and translate it to how it can impact the firm’s stock performance more effectively than how financial markets do today?
We will see enterprise-level copilots marketed sooner than businesses’ own readiness to adopt technology that cuts across their internal siloes. When they do, GenAI promises unprecedented integration of business.
4. Integration of Industries
Ask which industry is the biggest contributor to pollution and we get a lot finger-pointing: Agriculture blames manufacturing; manufacturing blames transportation, and so on.
Ask which industry in healthcare sector is driving up costs. We get similar answers: Health services providers blame health insurers; insurers blame pharmaceutical firms and providers; all blame the government…
While we may not get universal agreement on the root-causes of issues in environment or healthcare, we can definitely raise the level of debate with foundational knowledge integration across industries in a sector or across sectors.
Integration of Everything
Putting it all together, we get this possibility.
Mother-of-all-integrations of human knowledge — technology, knowledge domains, business, industries, and our entire society.
All these possibilities are tantalizing, rife with potential for abuse. Realizing that promise reliably will require us to get beyond technology that slices and dices information and constructs specious string of likely words. Even then, the mere fact that such scale of information integration is possible is exciting.
More Q&A
I went back to chatGPT and upped the ante. I added anthropology to the three fields — neuroscience, behavioral economics, and public policy — I asked about earlier.
Its answer was on the same lines. Summary of the response is here. [Full response from chatGPT is in Note 1 below.]
“In summary, these disciplines are linked by their shared interest in understanding human behavior, and their collaboration can lead to more holistic and effective approaches in shaping individual and societal outcomes through well-informed public policies.”
How often do we see multidisciplinary experts come together to solve societal issues as outlined in the statement above?
2023 maybe the peak of GenAI hype. Thoughtful critics from language experts to neuroscientists call it “autocomplete on steroids”, “another snow plough”, “automation of disinformation” etc.
While the current state of AI may not please our quest to reproduce human intelligence let alone consciousness, the promise of unprecedented integration of human knowledge and the potential for us to reimagine our world are palpable.
I consulted chatGPT’s own opinion on it. I got past its standard disclaimers and modesty of “I am only an AI language tool…” by flying under the radar. (Prompt engineering?)
That is a pretty good summary of this article itself. Good enough for SEO for sure.
It would have been self-aggrandizing if I asked it to replace “Generative AI” in that statement with chatGPT. I did not try, for I was sure it would not hesitate.
Notes:
- The intersection of these fields [anthropology, neuroscience, behavioral economics, and public policy] is particularly evident in areas such as health, education, finance, environmental conservation, and social welfare. For example, understanding the neural mechanisms of addiction (neuroscience) can inform strategies for combating substance abuse (public policy), considering cultural perspectives on education (anthropology) can improve educational programs (public policy), and recognizing behavioral biases in financial decision-making (behavioral economics) can shape regulations to protect consumers (public policy).
