In our Identiverse 2021 talk, we said AI is making healthcare’s Identity of Things more complex. “Ever smarter devices create gaps between identity capabilities of the latest devices and older ones.” Let’s expand a little.
Artificial Intelligence will be a lot more pervasive by 2030. There’s been a lot of work at bringing smarts into devices, buildings, vehicles, clothing, and the world at large. As the capacity for compute and connectivity increase, developers have been successful at squeezing more machine learning (see TinyML), speech, biometrics, and other pattern recognition apps into edge devices. So expect…
Rugged AI. Much of what we do in AI today assumes abundant power, compute, storage, and connectivity. IoT will increasingly require support for intermittent power and connections, and fixed on-device computing as devices are widely used in the developing world, in the great outdoors, and in military applications.
- The identity stack also must work in rugged environments. So, rugged identity.
Identity Inside. A Tesla vehicle has many subsystems, each becoming smarter individually, discovering better ways to coordinate and collaborate with each other. Bots are made of bots, and make up bots. Like the Theory of the Firm (companies add capabilities inside when things cost less than buying outside the company).
- Components may be coming from many vendors, and each message within the system must trust the other systems, and their parts.
- Each component may have the power to speak on its own to services outside the device. Like your pacemaker having a conversation without your consent.
Social Graph of Things. Devices will do many things humans did in social media. Discover each other, share information, coordinate work, build reputation and trust.
- The digital identity models of the 2030s will compose each thing’s identity from multiple sources. Trust emerges from the many relationships a device has with services, devices, legal entities, and the people who touch and are touched by the device.
AI’s shopping for themselves. If you’re running short of compute power for a problem, maybe you rent some on the open market? Been running on battery for a while and need a top up from a neighbor in exchange for an IOU? You’ll see devices making markets, buying and selling data and other resources from services and other devices.
- Decentralized identity hosts the trust that makes markets and commerce possible.
Identity conversations between people and devices. Healthcare, clinical and biopharma research, and regulatory bodies are demanding orders of magnitude more auditability and transparency of IoMT’s identity layer. The quantity, diversity, and novelty of this data overwhelms human cognitive capacity. [see also: “We need to talk…” UX for Identity of Things conversations]
- Pattern analysis and natural language processing will assist with digesting data by and about a device’s identity, security, data provenance, jurisdiction, ownership, custody, and GDPR roles.
- AI will also enable conversational experiences, explaining the overwhelm, simply, to human subjects, patients, caregivers, technicians, emergency personnel, and clinicians.
So: rugged AI (with rugged identity), congregate identity (things in things), social identity, identity for decentralized IoT commerce, and HCI for AI with a focus on identity conversations.
P.S. We walked through this and a few other ideas at Identiverse 2021. IoMT At Risk. A Wider Team Critique of Digital Identity Threats to the Internet of Medical Things. Read all the posts from our Identiverse talk.