Let’s envision a futuristic scenario: You’re hiking and find yourself lost in the woods, Hansel and Gretel style. You call on a robot to rescue you. It can’t just be any old robot, though, but rather a trusted robot, one that you feel confident will get you to safety. This robot also has to trust you—it’s not saving you for free (#EqualPay for robots), and so it needs to know that it will get compensated for the task of leading you out of the woods.

This may not seem like the most plausible scenario. If you’re trapped somewhere and have the cell service required to summon a robot, you’ll probably just call 911 instead. Human-led emergency services will get your GPS coordinates and come save you, hopefully.

But this is the use case for blockchain-controlled robots that Intel demonstrated during Mobile World Congress last month. Using the Enterprise Ethereum Alliance’s Trusted Compute API (TC API), Intel—along with iExec and ShanghaiTech University—was able to showcase a scenario in which a blockchain-controlled robot “rescued” two, stranded robots and led them back to “safety.” You can watch the video below.

Intel has been collaborating with multiple companies in the EEA to work on better connecting the Ethereum blockchain to items not directly attached to a blockchain since at least October 2018. That was when Intel’s blockchain program manager Michael Reed announced the Trusted Compute API. The EEA members working on the TC API include ConsenSys, Enigma, Microsoft, and Oasis Labs—in addition to Intel and iExec. Their aim is to bring blockchain to the growing Internet of Things by not only applying smart contracts to off-chain assets, but also establishing “reliable links” between the blockchain and those physical items—connections that have been notoriously difficult to establish.

So how is the group attempting to, as Reed puts it, “extend decentralized trust to off-chain assets”? Let’s break down the process. First, a marketplace smart contract is established on the Ethereum blockchain (it could be any blockchain, but the group is working on Ethereum since it’s part of the EEA). The smart contract registers service-providing robots along with the services they perform (like AI navigation services, in the case of the demo). If another robot or a human is in need of a specific service, they can scan the smart contract to find the nearest robot able to perform the task.

Get the BREAKERMAG newsletter, a weekly roundup of blockchain business and culture.

The contract then “conveys a going rate for service based on the supply of capable robots and [the requester’s specific] demand,” explains Reed. The nearest, most capable robot is found to carry out the task. Once an event takes place that has been predetermined to signal the task’s end, the smart contract logs it as “completed.” Then the robot—or rather, the robot’s operator/owner—gets paid (in this case, in ether).

“The reason we built this demonstration with iExec is to show the interaction of IoT devices with the advent of 5G, which was a very popular topic at Mobile World Congress,” Reed tells us. “The low latency and the improved bandwidth of 5G suggests that IoT devices are going to grow in their sophistication.”

This means more, new use cases for internet-connected technology. Some of the examples Reed offers are drone delivery and, perhaps even more seemingly far off, remote surgery, where a skilled surgeon who lives across the globe can instruct operating tools at your local hospital. “I honestly don’t know which will come first [drone deliveries or remote surgeries], but they’re both futuristic,” he says.

A more imminent application might have to do with self-driving cars. Autonomous vehicles are full of sensors that tell the car, for instance, whether it’s in danger of hitting something or veering into another lane. Reed asked me to imagine a “fleet of autonomous vehicles” connected like the robot rescuers in the Intel and iExec demo. Living in Portland, Ore., Reed often has to drive through mountain passes, which can be dangerous in bad weather. “In the future, I might be able to get a real time video of the mountain passes” that reveals road conditions, he says—one taken using the video camera on an autonomous vehicle and authenticated using an RFID chip connected to a blockchain smart contract.

Related: Can Aragon Build an Unstoppable Robotic Government?

Self-driving cars may be less far off than remote surgeons, but they’re still full of problems that prevent them from becoming mainstream. Consider the reports coming out that autonomous vehicles are racially biased and have learned to make fatal decisions. What other issues could plague connected smart cities, where blockchain-based smart contracts might help us find, request, and compensate our ubiquitous fleet of robot assistants? One may be inclined to imagine the worst—scenarios in which people plant weapons in drones they know are headed to specific locations or hijack the network for nefarious purposes.

“Wow. So that’s obviously a fairly grim perspective,” says Reed. He insists that the TC API is being built with trust and reliability in mind. “The point you’re bringing up is one of the reasons Intel and several other developers are working on building a programmatic interface to off-chain assets—to try to improve the reliability of those assets.” In other words, it’s the verifiable connections between the blockchain and the IoT objects that should allow for more trust in the latter. Still, he admits, “Anything is possible.”