I am writing you on a proposed method to help increase the contribution to reducing global warming on an individual level by creating a planning aid which helps to inform the user about climate-friendly behaviors. We have seen how irrational many people behaved in the face of the pandemic, but also many people did not. Perhaps using normative methods, we could encourage widespread adoption of the system I am about to describe, working title - the "Free Life Planner." The system would help people to act responsibly, by informing them intelligently about their present behaviors, and suggesting modified behaviors to achieve their objectives. This system helps users through a "rule"-based system which assists them to tailor their planning, scheduling and execution of tasks in line with better methods of helping the planet. For instance, if the person was processing their trash incorrectly, this system would help to educate them. The best thing about this system is that it consists of no additional hardware, only free software (which is in an advanced state of development currently). What is lacking is not the software, but the social capital possessed by leaders like you to bring it to fruition. The basic idea is to piggyback off a set of physical sensors in smart phones and smart homes, and collate their data into a central subsystem called the "World State Monitor (WSM)." The WSM will act behind the scenes, the users won't be aware of it. The WSM is based on semantical assertions, like: + "Person 1 is approaching the trash can." + "Person 1 is holding the cell phone." The FLP would also contain facts such as: + "FLP has never told the user about proper recycling methods." but formatted in a "rule"-based logic-based computer language (currently Prolog). The system has the capability using logic programming to draw the conclusion that it should try to inform the user about their localities. Here is a sample rule I am writing on the fly to illustrate. tell(flp,Person,proper(methodsOf(recycling))) :- approaching(Person,trashCan), holding(Person,CellPhone), not(told(flp,Person,proper(methodsOf(recycling)))), This poorly constructed rule in theory would schedule an immediate action to inform the user about the proper method of recycling when they were approaching the trash can, holding the cell phone, and had never been told about it before. This was not the best possible example since we would have to presume image understanding routines to locate the trash can, but such cases are not necessary for the system to work well, there are plenty of other ways in which it can be successfully used. In general the WSM tracks a multitude of facts about the person's situation. It also uses planning, scheduling and execution algorithms to interactively walk users through tasks. Over 20 years of work has been conducted on this system, and you may read more about it here: http://freelifeplanner.org I will now give two more on the fly examples in order to try to make the point more clear. Suppose the fact was observed that: + "Person 1 is leaving their house." - from cell phone geoposition data (already works btw) + "Person 1's heat is on." - from the smart thermostat. + "Person 1 lives alone." - from social media, email correspondence, user knowledge elicitation dialogs, etc. You might have a rule like: ask(flp,person1,permissionToLowerThermostat(Home)) :- livesIn(Person,Home), likely(unoccupied(Home)), leaving(Person,Home), holding(Person,CellPhone), hasTemperature(thermostatFn(Home),fahrenheit(TemperatureInFahrenheit)), TemperatureInFahrenheit > 70. as well as the rule: likely(unoccupied(Home)) :- livesAlone(Person), owns(Person,Home), not(atLocation(Person,Home)). This previous use case is probably already covered by most smart thermostats I'm guessing, but it illustrates the principles involved. What we need to do is develop good rules that people should do. I have systems under development for taking texts like guidance from ready.gov and so forth and converting it to the proper formats for emergency preparedness. I guess that leads naturally to another example, so I will give one last example of this smart FLP system: scheduleTask(flp,Person1,checkBatteriesOnAllFireAlarmsInHome(Person,Home)) :- lastTime(happened(Time,checkBatteriesOnAllFireAlarmsInHome(Person,Home))), currentTime(CurrentTime), duration(Time,CurrentTime,Duration), Duration > months(6). Here the system is reminding the user to check their fire alarm batteries. It might combine that with locality specific advice for getting rid of the batteries. The basic principle is people might be more willing to act in ways conducive to stave off global climate change if they are educated on the spot, contingently, on the basis of what the Artificial Intelligence knows about their situation, and vast sets of rules that hopefully will be developed by responsible parties. This is just the beginning though, there are sophisticated research-grade A.I. planning, scheduling and execution systems already integrated that combine this action guiding capability (to inform the user of better methods) to achieve their objectives. Help us Obi-Wan Kenobi, you are our only hope!