Ambient Intelligence Needs Better Signals; Immersion Already Has Them

OpenAI CEO Sam Altman and designer extraordinaire Jonny Ive just announced they will be collaborating on a device to identify when people need an AI assist.  There are many open questions about how such a device might work, but importantly, such ambient intelligence is a signal extraction problem.  By that I mean that given all the things people are doing all day, when should a trained AI model send you information and when should it hold back and let you just be in the experience or for that matter just rest?

There are several ways an ambient AI could be trained.  One would be to randomly ask people questions about what they are doing.  This is known as the experience sampling method and it has two flaws.  First, such questions interrupt people during the experience they are having.  If the experience is important, or enjoyable, or exciting, the interruption has just ruined it.  Second, if you ask people to rank how important their present experience is, the signal extraction is anchored to self-reported feelings.  As I have written about before, feelings are our conscious evaluation of our unconscious neurophysiologic states and they are seldom accurate because the brain is not designed to report its inner workings.  Yup, Freud was wrong about bringing the unconscious into consciousness (he also consumed cocaine, morphine and opium among other substances so let's use modern neuroscience rather than ungrounded speculation from a long-dead druggie).

Any alternative approach, and the one my team at Immersion Neuroscience has pioneered, is to continuously measure the value the brain obtains from experiences.  This is precisely what Immersion's SIX app does: measure the neural value of experiences every second all day, cataloging where and when the most valuable experiences happen.  These high value Key Moments are experiences that people care about the most.  This is when an AI assistant can provide additional insights so people can get even more value from what they are doing. 

On the other hand, when neurologic Immersion is low, a person is in "idle" mode and is chilled out.  The brain needs time to recover after working hard, just like muscles do, and a poorly designed ambient AI would try to fill these downtimes with information.  This is a bad idea. Immersion data have a wave pattern: Immersion rises and falls, and Immersion peaks are always followed by troughs.  Long extended peaks are highly valued experiences, while troughs allow the brain to recover and reset so it is ready to have its next adventure.  Indeed, the recent films featuring elite athletes in Toyota trucks that used Immersion's SIX app to measure adventures neurologically showed exactly that: Key Moments in the brain could be sustained for a sustained period of time but then brain cells need to rest to prepare to reach the next Key Moment.

Immersion data are continuous, objective, and accurately predict behaviors.  SIX is neuroscience on your arm.  This will shock you, but the brain is connected to a body and they talk to each other.  The SIX app uses cloud-based algorithms to capture this communication channel from the cranial nerves. It seems like magic--it just took our team 25 years of published scientific advances to make this simple for users and highly predictive of users' needs, wants and behaviors.

OpenAI is not the only company developing ambient intelligence; other major companies in this race include Microsoft, Google, and Amazon as well as startups like Anthropic, Brilliant Labs and Humane (now owned by HP).  At least so far, OpenAI seems to be the furthest along in wearable tech scheduled to launched in 2026 after Humane's voice-operated "pin" failed.  Other approaches use IoT devices to prompt ambient AI.  The beauty of wearable tech like the SIX app is that users can turn them off.  When IoT devices are on and working, users are often are unaware of them and may not know how to cycle them down.  The SIX app is on users' smartwatches.  The data stay private to the individual unless they choose to share it, and it is easy to simply turn off one's watch.

SIX is ambient intelligence. SIX measures the primary not the proxy; that is, it measures what the brain actually values and desires more of; the is much more effective than trying to infer this from behaviors or self-reports.  It is only a matter of time before SIX, which already has a trained AI assistant, is used to train your own private LLM. 

SIX is free and fun, and every day gives users insights into their true selves: What they love doing, who they love doing it with, and what they can do more of in order to thrive.  Ambient intelligence is here, try it today.