Science and Technology can improve our quality of life; S&T Policy is a method governments use to enable that improvement. The recent OSTP report Science: A New Golden Age (SNGA) reexamines America's research and development ecosystem in light of changes brought on by the 21st century. It identifies four overarching goals. This note takes up two of them.
The role of a builder-philosopher is to design and develop technology that improves lives and enables "the good life" as envisioned by each person in their own way. Drawing on the philosophy of engineering and technology, this note argues that SNGA's critique of scientific and innovation processes fails to recognize that role in securing the pursuit of happiness. The report needs to better consider socio-technical aspects, and the role of the Federal Government in addressing them.
Technology mediates the relationship humans have with the world (Verbeek, 2005), and that mediation shapes experience and existence. In our age, S&T plays an outsized part in defining the methods by which a person can pursue their own good life — which makes the design of technology a moral matter, not only a commercial one.
"Scientists discover nature's secrets; Engineers create new order."
The rubric SNGA's own examples support
SNGA critiques Vannevar Bush's linear sequence of discovery and innovation as no longer adequate, claiming the model is broken because discovery now happens in industry. But its examples only show that innovation in industry is possible because of discovery in science.
The transistor was discovered and first implemented at Bell Labs — a lab with steady funding that made research for the public good part of its reason for being. The AI breakthroughs came from Hopfield and Hinton as university researchers, later awarded the Nobel Prize, years before the LLM revolution. The transformer architecture and the integrated circuit were engineering innovations built on those discoveries.
The feedback between discovery and innovation is real, and SNGA describes it accurately. But it reinforces the relationship rather than refuting it. What is called for is refinement of that feedback — and recognition that innovation has always been better in industry.
Four Topics
From the builder-philosopher's place in living well, through design as the creation of order, to the balance of freedom and risk and the demands AI places on both.
Where science discovers truths about the natural world, engineering creates entities that never existed before — and that creation places a moral onus on engineering. Engineers exercise significant, often tacit, power in shaping societal infrastructures (Polanyi, 1966; Vincenti, 1990). S&T Policy must decouple discovery from the push for innovation, and support engineering innovation in industry more deliberately than ever.
Design achieves function through physical implementation; balancing the two is an ethical act as much as a technical one. Aristotle observed that it is the mark of the wise to order; Aquinas reformulated it as sapientis est ordinare. Simon made design the proper study of mankind; Alexander charged builders with aesthetic-moral responsibility for the wholeness of what they build. Practical wisdom belongs inside engineering design, not beside it.
We need not fear discoveries of nature's secrets, but we must be hesitant when creating new order. SNGA frames innovation as permissions — an odd framing to an engineer, since permissions are elements of regulation, not of engineering judgment. Freedom to innovate must be balanced against the risk innovation creates, and that judgment belongs in development, where risk management already lives.
AI is redefining the science process, our understanding of the natural world, and the practice of design itself. SNGA seems to imply human practical wisdom matters less as cognitive capability becomes abundant. The opposite is true: in the age of AI we must redouble efforts to guide science and technology by human society considerations.
Critique
It confuses private-sector development with research
Citing the transistor and the transformer as products of "corporate laboratories" conflates the discovery that made them possible with the innovation that brought them to market. Technology development has always iterated between the two.
It leaves the social sciences out of innovation
SNGA calls for social technologies to improve the R&D process, but never for research on social technologies to improve science and technology themselves. It critiques psychology's reproducibility crisis without describing what psychological and other social research offer a new golden age.
Permissionless innovation misplaces the judgment
When permissionless innovation happens in industry, the overwhelming factor in risk assessment is profit, with negative impacts postponed. "Move fast and break things" shifts risk from the developer to a consumer unable to assess it. Engineers are ethically charged with weighing the balance; corporations have shifted that assessment to other parties.
Regulation is treated as the roadblock
Assume instead that regulation aims at the social good. Then S&T Policy should emphasize discovery and innovation that make technology impacts easier to assess — making compliance a normal course of innovation. SNGA's observation that talent migrates to where regulatory burden is lightest should be read as a call to restore the engineering role of ethical design.
We have blinked at this revolution before
The benefits of social media were claimed with little support by the industry building it. The absence of socio-technical consideration led to addiction and harm — and, eventually, to large penalties and platform redesign. Compare the Human Genome Project and SEMATECH, where pre-competitive consortia coordinated innovation before market pressure determined the outcome. We can do better.
The AI Turn
AI brings new perspective to what we mean by agency. Agents sense and respond to their environment and act toward goals; we have long studied agency in the world, and now we are creating it in earnest — which is changing our understanding of the world. AI has redefined cognitive space itself through the spatial embedding of concepts, and the engineering of autonomous agents in cyber-physical ecosystems must include philosophical considerations (IEEE 2874 Spatial Web).
Earlier phases of the AI revolution were paired with inquiry into how the computing aspects of AI relate to the human aspects. That research program was cybernetics, and the cybernetic moment ended as cybernetics was flattened to the all-purpose adjective "cyber" by the 1980s (Kline, 2015). Those socio-technical topics — now spread across the natural and social sciences — need to become a primary point of emphasis in S&T Policy for AI.
Dennett identified the real danger of AI: that we over-estimate AI tools and prematurely cede authority to them. The tension between freedom and risk in AI design and innovation is perhaps the defining element for S&T Policy in our time.
In Closing
Yes to both of SNGA's goals as stated — and in each case, further than the report goes.
SNGA Goal
"The Federal Government should set clear scientific goals and build the industrial muscle to translate scientific discovery into technological strength."
Yes, but US S&T Policy must go further. It must clarify the difference between scientific discovery and technology innovation. With that clarity, the role of engineering design and the balance of reward and risk in innovation become clearer. As the engineering of innovation grows more bound up with socio-technical ecosystems, research and innovation in the natural and social sciences is needed.
SNGA Goal
"We must prepare our research enterprise for the AI revolution. AI will accelerate and radically transform the way we do science."
Yes, but we must not lose the role human wisdom plays in defining that iterative path of AI discovery and innovation — the path that gives everyone the opportunity to define how technology serves what they consider the good life.
Related Work
Interested in the socio-technical foundations of S&T policy, engineering ethics, or AI governance? Get in touch.
✉️ percivall@ieee.org