Five federal agencies/entities are driving three presidential nanotech-based initiatives that coalesce into a single challenge: “develop transformational computing capabilities.” This single challenge is divided into seven core R&D focus areas, one of which I have selected to serve as the focus of this post: “Brain-Inspired Approaches.”

All of this is the construct of yet another new ecosystem of magic-like computing power. When taken together with quantum computing, which has already produced viable quantum computing states and 3-bit cell in-phase change memory, the picture becomes clear that this process, this “Transformative Computing” effort, is driving more than a mere paradigmatic shift. It will perpetuate an epistemic break.

The allure of replicating the brain’s capacity to compute, learn, adapt, physically grow and regenerate in the form of AI applications is (rightfully) irresistible. Throw in the fact all of these operations are performed using little electric power, about 20 watts, makes it even more so as the possible applications are considerably magnified.

Be it through reverse engineering efforts (neuromorphic computation) or artificial neural networks focused on machine and deep learning, Transformative Computing will contain components that mimic, perhaps even exceed, human brain capacity. We are already witness to the design of artificial neurons that use phase-changing, binary state (amorphous and crystalline) materials (germanium antimony telluride). Binary-state ions housed in fluid nanoelectronics/nanoionics also offer a hint as to how the proto applications will function. Development of these types of neurons is critical to brain function replication. They have the capacity to perform the necessary complex computational operations and have brain-like power consumption characteristics. This, of course, is a key feature, especially (albeit not exclusively) for wetware-based Transformative Computing. It also dovetails with Kurzweil’s prediction that within 30 years human intelligence will multiply billionfold once our neocortex wirelessly links to cloud-based “synthetic neocortex.”

So where does all of this fit with the law? For starters, it is reasonable to conclude that the law will not proactively evolve to address Transformative Computing. There is no relevant precedent to support an alternative prediction.

But here’s the twist: The very thought that the law can, needs to, or even should be proactive to such dramatic changes is one plagued by flawed legacy logic. The computational power that will be present at the dawn of the Transformative Computing age stands to vaporize contemporary legal analytical frameworks. “Just” considering the prospect of widely-distributed nano-level autonomous AI capable of efficiently mining through exabyte (and larger)-sized data sets, learning, building and quickly executing decisions, teases us towards a genuine reconsideration, if not acceptance, of what law making will look like 30 years from now.

Transformative Computing computational law applications can be expected to play a significant role in this. Consider, for example, this cloud-based synthetic neocortex link, enabling a slew of wetware-based AI computational law applications. The massive, ubiquitous, bi-directional data feeds that contain logically-structured (read well-thought) legal analysis, are poised to fuel “organic legal growth.” This concept/prospect is equally exciting and promising.

Think about it: an organic, exadata-fueled legal regime that is readily accessible and understandable to the masses may be just 30 years away.

***

Updates & Additional Thoughts

4/16/19: Frontiers in Neuroscience recently published the “Human Brain/Cloud Interface” (B/CI)) paper written by UC Berkeley and the U.S. Institute for Molecular Manufacturing. From the description of what the researchers call “neuralnanorobotics,” it is evident that they are basing their work on the brain-inspired approaches discussed above, including Kurzweil’s synthetic neocortex. The researchers describe a brain interface where the neuralnanorobotics “…wirelessly transmit up to ∼6 × 10¹⁶ bits per second of synaptically processed and encoded human–brain electrical information via auxiliary nanorobotic fiber optics (30 cm³) with the capacity to handle up to 10¹⁸ bits/sec and provide rapid data transfer to a cloud based supercomputer for real-time brain-state monitoring and data extraction.” When you consider this together with what I wrote above, it becomes evident that the groundwork for Transformative Computing is expanding. This, in turn, further fuels the urgency around implementing appropriate security and privacy-design practices, along with an effective legal regime that will help ensure not only end-user safety, but also privacy. This view also syncs and advocates for the proposition that the B/CI ecosystem should be viewed as a cybersecurity threat surface. Once that conceptual and operational framework is established, it becomes clear that tools such wetware-based AI computational law applications, (neuralnanoapplications) can have a critical role to play in making sure that threat surface is properly maintained.  (See also, “Rise of the Intelligent Information Brokers: Role of Computational Law Applications in Administering the Dynamic Cybersecurity Threat Surface in IoT” 19 Minn. J.L. Sci & Tech. 337 (2018)).

 2/13/18: In terms of the cognitive-type analytical functions (see 2/11 update below), exascale computing could enable sophisticated contractual nondisclosure and fraud analysis. Here we see an AI-enabled computational law app making real-time decisions on whether delivering less information is more likely to prevent an instance of misrepresentation for a particular consumer (based, in part on the user’s acceptable risk profile).

02/11/18: Accelerating the proliferation and power of deep learning algorithms is dependent on inexpensive access to high-speed computing platforms. Google’s Cloud TPU architecture is publicly available, enabling access to 11.5 petaflops. Of course, the quest for faster, inexpensive platforms will inevitably surpass the current leader, China’s Sunway TaihuLight (93 petaflops) and venture into exascale computing, which perform (at least) a billion billion (exaFLOPS) calculations per second. With this magnitude of computing power, big data analytical capabilities are exponentially magnified, enabling sophisticated computational simulations that are currently unattainable. In this environment, AI-enabled computational law applications should be able to execute cognitive-type analytical functions, which could assess various complex business transactions quickly and efficiently. As such, we will see these CLAI apps initially emerge as business applications, which will, over time, make their way into consumer-level apps.

5/24/17: Ensuring data integrity will be the holy grail in the era of Transformative Computing. The blockchain will play a critical role in this respect. To be feasible, the “meta-trust” principle first needs to be satisfied. Within the conceptual and operational framework of the blockchain, meta-trust means the blockchain meets the structural and operational features that render it immutable. One the blockchain meta-trust standard is satisfied, it should be formally represented in the blockchain implementation in the form of certification. Because certification signals compliance, it serves as an essential catalyst for breeding trust, which drives blockchain adoption and enables (from a practical perspective) the drafting of formal rules (laws and algorithms) that govern user and application activity.

2/28/17:  Can the concept of a cloud-based synthetic neocortex be seen emerging from the collaboration between Stanford’s Laboratory of Behavioral and Cognitive Neuroscience (LBCN) and the Stanford Human Intracranial Cognitive Electrophysiology Program (SHICEP)? Possibly. Though the immediate application aims at controlling a prosthetic, similar principles could (arguably) be employed to build the synthetic neocortex link.