Treatment Using AI/Data Science to help treat addiction

[ownstrength]

I've been a software engineer for a few years and in that time climbed my way up to FAANG, shorthand for the top 5 tech companies altho that's debatable. I am skilled with CompSci fundamentals and effective at passing technical interviews, but I lost nearly 20 years of my life to active addiction. I had to do less mentally demanding work, but still did some good. I feel if I am able make a contribution to the recovery field in my life will not have been wasted in vain. I am Buddhist, and considering some meditative approaches, but they're very challenging and I believe "God helps those who help themselves. esp with data science n stuff."

I had several clean years but one thing led to another and attaining to the higher levels of my dream job was overwhelmingly anxiety inducing. and I relapsed just before the mass firings, they just sealed the deal as I was junior and near the top of the list so I was fired too. My performance was fine I was assured, and I know I was able to solve problems my predecessor who moved to another tea couldnt. So I may not be brilliant, but I'm good enuff. I see a few ways to use the "advanced AI" / data science to assist with the problem. For example. In my case. I am sure a well-trained AI could see my relapse coming from months away and avoid it, kind of like keeping asteroids from hitting earth by using a weak gravitational force to change its angle minutely.

Then ways to help the addict in active addiction get out of it by various ways such as empowering them with awareness, predicting the spikes in their cravings, optimizing MAT schedules and methods, and so on. I have some more specific ideas but wish to retain the IP as I am no greedy but wish to use the resources for the wellfare of the many, unlike most companies.

The purpose of my post is to discuss the hardest part of this process: getting the data. Between privacy laws and personal issues, this is a major challege to navigate, and esp with limited funding. A big player COULD do it, but if there were sufficient profit in it, it would serve the private corporations first and foremost. I am a unique player, very convinced of karma, and believe I could abstain from corruption by wealth and prioritize the people.

For active addiction management, might need wearable components to collect enough data about bodily systems. Not sure if it could be done via phone alone.

Thoughts I have had:
1) approach recovery centers, public and private, have other useful features in app to encourage adoption. Those that partner with app and allow it to collect data will get cheap share in benefits.
2) State/Federal funding: since this is becoming national level problem.
3) See about buying some from China, and mining in other societies w/ less privacy
4) Transfer learning, what mechisms have similar nufff dynamics for this to work?

PS
Does the paranoia of actiive addiction make this idea sound evil? Modeling our minds amd behaviors will continue to develop. Might as well have some of it, empower the people, and to use for their benefit

 

[someguyontheinternet]

Definitely raise an eyebrow whenever someone says they want to collect data. Is it covered under HIPAA? How private is it really? There is already a stigma against drug users, would this make someone's life worse if that data were to get out? Also a bit suspicious that you suggest "mining" in societies with less privacy. Is that a good or right thing to do? Is that exploitative?

If you want to help using computers, I think the most effective way right now for data science is to do things like analyzing the literature looking for drug interactions or analyzing genomic data for bioinformatics or proteomic analysis. Getting a better understanding of the mechanisms of mental health issues has a massive impact on the future treatment of addiction because many people with addiction issues also have issues with mental health in some fashion. Computers can be used to accelerate research and analysis to a large degree

 

[ownstrength]

Many already say that privacy is a very relative perception, hardly existant. The companies privvy to "Big Data" are likely to already be able to easily determine if a person's life is dominated by addiction. Stigma wise, I do not think biometric and behavioral details (in terms of cravings, usage) it could get worse thru this.

"Is that a good or right thing to do? Is that exploitative?" The first question would depend on the intention of the actor. In the past year I revealed my addiction on facebook, which is close to official. If it is helpful I will do so on linkedin. My intention is to attempt to reduce to suffering and restore agency to people's lives. I do not think it's exploitative to function within different areas of an international community legally to achieve this.

I'm all about the mental health aspect. understanding that nobody chooses to become an addict. and that full recovery requires activities, housing, employment, community, etc.

But the behavioral state of active addiction has to be interrupted. And if data science can make that orders of magnitude more likely it should. The problem is that an addict has by and large already lost agency, are paranoid, and by mental unhealth clinging to addiction. Even mental health problems can be massive treated by data science.

The universe functions on laws, not philosophy. I'm not sure that privacy, esp the hiding of unhealthy things is considered of value to reality/life/the universe.

It would have to be legal, for sure. But for example Ive been to private rehabs with lots of cameras collecting data if you want help. I am inspired to do something becuase THEY are exploittaive, there is a huge private recovery industry in the US that treats people like shit and has shit outcomes. I'd like to improve lives so I can die feelings better about my contributions. Whether others believe it or not, those are very different intentions.

 

[someguyontheinternet]

The universe functions on laws, not philosophy. I'm not sure that privacy, esp the hiding of unhealthy things is considered of value to reality/life/the universe.

Revealing an addiction to state actors can be unwise for a person's health and wellbeing depending on where they live. How do they treat addicts in the Phillipines? Are there other regimes that impose jail time or death on addicts? You mentioned China, how do they treat addicts?

 

[ownstrength]

I agree with that. I certainly would not reveal this data to state actors like philipines. I named China because they already have been generating total-surveillance data including about addiction dynamics of their citizens. Off the top of my head I dont know how China treats addicts, do you? I have an engineer friend who worked there for some time and says they employ some pretty strange methods with the recovered addicts walking out of the "centers" smiling. One might argue that it's not inferior to most of the world, which has a shit reaction to addiction except for the few countries who have the political will to act on the data and do radical things like exterminate the black market by putting people on heroin maintenance. I cant see the US being capable of doing "the right thing". Highest cause of death for 18-35 males is OD here, largely due to fent but addiction is growing and the horrid power of synthetics drugs is too. We need a heroic solutution, racing on electric fields and increasing accuracy with sample set size. it's going to take a data scientific superpower to save us (trying to resonate a song).

But I would say China's methods are too brute force. I'm into elegant solutions, and all they require is sacrificing the illusion of privacy or that it's worth destroying one's life for.

 

[ownstrength]

The argument can also be made, and has, that people whose frontal lobes are conquered by addiction have lost their agency. That's part of the compassionate view. I tell people drug addiction is like AIDS, it attacks the parts of the brain responsible for defense (executive functioning, mindfulness, impulse regulation.) Its either that or a "moral failing" like "you just lack the human will to recover, you are devoid of worth." Of course that is meant facetiously.

 

[Jabberwocky]

Interesting idea, but as you say, the problem is getting the data.

I wonder if an app could be created that analyzed your device usage patterns and warned you if it suspects you might be showing relapse behavior? For example, if someone was in recovery and did not want to relapse, an AI app could analyze their usage patterns/search history, sleep patterns, or other things, and then warn the person "you may experience greater than usual cravings or might relapse soon"? Not sure how practical or useful it would be, but it got me thinking.

 

[someguyontheinternet]

You say you wouldn't reveal info to state actors but that's far easier said than done. How would you accomplish this? Where is the data stored? Is it E2E encrypted? At rest on your server? Who has the keys? Can you resist legal orders?

 

[SpiralusSancti]

I've been a software engineer for a few years and in that time climbed my way up to FAANG, shorthand for the top 5 tech companies altho that's debatable. I am skilled with CompSci fundamentals and effective at passing technical interviews, but I lost nearly 20 years of my life to active addiction. I had to do less mentally demanding work, but still did some good. I feel if I am able make a contribution to the recovery field in my life will not have been wasted in vain. I am Buddhist, and considering some meditative approaches, but they're very challenging and I believe "God helps those who help themselves. esp with data science n stuff."

I had several clean years but one thing led to another and attaining to the higher levels of my dream job was overwhelmingly anxiety inducing. and I relapsed just before the mass firings, they just sealed the deal as I was junior and near the top of the list so I was fired too. My performance was fine I was assured, and I know I was able to solve problems my predecessor who moved to another tea couldnt. So I may not be brilliant, but I'm good enuff. I see a few ways to use the "advanced AI" / data science to assist with the problem. For example. In my case. I am sure a well-trained AI could see my relapse coming from months away and avoid it, kind of like keeping asteroids from hitting earth by using a weak gravitational force to change its angle minutely.

Then ways to help the addict in active addiction get out of it by various ways such as empowering them with awareness, predicting the spikes in their cravings, optimizing MAT schedules and methods, and so on. I have some more specific ideas but wish to retain the IP as I am no greedy but wish to use the resources for the wellfare of the many, unlike most companies.

The purpose of my post is to discuss the hardest part of this process: getting the data. Between privacy laws and personal issues, this is a major challege to navigate, and esp with limited funding. A big player COULD do it, but if there were sufficient profit in it, it would serve the private corporations first and foremost. I am a unique player, very convinced of karma, and believe I could abstain from corruption by wealth and prioritize the people.

For active addiction management, might need wearable components to collect enough data about bodily systems. Not sure if it could be done via phone alone.

Thoughts I have had:
1) approach recovery centers, public and private, have other useful features in app to encourage adoption. Those that partner with app and allow it to collect data will get cheap share in benefits.
2) State/Federal funding: since this is becoming national level problem.
3) See about buying some from China, and mining in other societies w/ less privacy
4) Transfer learning, what mechisms have similar nufff dynamics for this to work?

PS
Does the paranoia of actiive addiction make this idea sound evil? Modeling our minds amd behaviors will continue to develop. Might as well have some of it, empower the people, and to use for their benefit

It's not even a question if AI will help understanding/curing/lessening/whatever addictions but as for personal AI warning me to stop, idk, I don't even listen MyI. No need to be a rocket scientist to calculate how to avoid addiction, more so with countless before going trough similar or sometimes even with previous experience.
Also no need to be a rocket scientist to calculate if that crazy girl will cost a lot and MyI keeps screaming stop yet my heart and dick don't let go..

 

[ownstrength]

You say you wouldn't reveal info to state actors but that's far easier said than done. How would you accomplish this? Where is the data stored? Is it E2E encrypted? At rest on your server? Who has the keys? Can you resist legal orders?

I haven't seen data aside from some limited personal details like a sha'd password encrypted beyond whatever db settings are used at 2 startups and a big-5. It is constantly used. The security is mostly at layer 4 with an ssh certificate on a sever behind a solid firewalled env and as you say E2E encryption using HTTPS requests. Currently my server and dbs are on linode, but some stuff is stored in google. I am not prepared for legal or forceful resistance and do not see it as relevant because big data can identify an addict quite easily unless they are super technical and disciplined in their addiction and that's not most people. I'm talking about biometric data to train a model to gauge and predict craving level, the occurrence of usage/relapse. Trying to trace usage intention as early as possible. This is not data exposing anyone to realistic new threats. Any biometric data is usable for the development of harm but I would be a drop in the ocean. Data collection much deeper than this has been taking place. It's more a question of using some of it by intelligent actors for wholesome causes out of a compassion that only decades of. the humility of defeat can bring.

You ask good questions tho. When a worked a little bit at a FAANG company scraping web data from another company whom we had a contract with, so it was technically ok, but kinda unofficial, people were worried. I was like, everyone on this site can see this data, it's practically public. I'm just a lowly engineer who makes the bits go. The big companies make a certain show of compliance. The small ones hardly care. Nobody is auditing them, checking to see if they actually delete user data or just flip is_deleted in the db.

That's in the US, the GDPR of EU is one of the more protective regulations, maybe they audit startups there. Def not storing servers there.

I just want to take a super-technical assault on the Enemy that has more or less destroyed all our lives and free the people enslaved by it. No data I create will be more weaponizable or dangerous than what's already been done. But Instead of benefiting the few, i want it to benefit the many who suffer. When I was in active addiction (which I'm back in for now but retain some effects of clean time) I was very paranoid and had a much more adversarial view of things, but like I said, I'm just an addict who lost a generation to addiction. Maybe I have the brains/qualities (grit, discipline, organization, non-defeat) to react, probably not.

 

[ownstrength]

It's not even a question if AI will help understanding/curing/lessening/whatever addictions but as for personal AI warning me to stop, idk, I don't even listen MyI. No need to be a rocket scientist to calculate how to avoid addiction, more so with countless before going trough similar or sometimes even with previous experience.
Also no need to be a rocket scientist to calculate if that crazy girl will cost a lot and MyI keeps screaming stop yet my heart and dick don't let go..

I hear you but my understanding is that awareness can have profound effects. For example if every morning you'd get a graph of your cravings, it's likely your behavior would change. One. your conceit would be displeased that you are so predictable, and when you had desire to quit it would give you more resilience, "this is just the 3:15 craving as the graph predicts, I can get over this!"

And that's just pure awareness, then there's things the app could do to help you counter the craving peaks.


It doesn't have to be called a "Mindfulness-Based Trademarked Method" to work, but I asked GPT4:
Me: I recall reading a study (I BELIEVE it was re anger and migranes) that mere awareness can prevent a medical ailment from arising

El GPT4: Yes, there is indeed research that suggests mindfulness and awareness can have positive effects on various health conditions, including potentially reducing the frequency or severity of migraines. Mindfulness-based interventions, such as mindfulness-based stress reduction (MBSR) and mindfulness-based cognitive therapy (MBCT), have been studied for their efficacy in pain management and stress-related ailments.
The mechanism behind this could be related to the concept of neuroplasticity, where the brain's structure and function can be changed by experience. Mindfulness and meditation practices can lead to changes in the brain regions associated with attention, emotion regulation, and pain perception.
While it is not accurate to say that awareness alone can completely prevent a medical ailment from arising, there is evidence that such practices can improve coping strategies and reduce the impact of symptoms, including those of migraines. It is always advisable to consult with healthcare professionals alongside mindfulness practices for any medical condition.

 

[someguyontheinternet]

I hope you know that gpt4 regularly hallucinates references and mechanisms. I use it for my writing sometimes and its something I'm incredibly cautious of

 

[ownstrength]

I hear you but my understanding is that awareness can have profound effects. For example if every morning you'd get a graph of your cravings, it's likely your behavior would change. One. your conceit would be displeased that you are so predictable, and when you had desire to quit it would give you more resilience, "this is just the 3:15 craving as the graph predicts, I can get over this!"

Interesting idea, but as you say, the problem is getting the data.

I wonder if an app could be created that analyzed your device usage patterns and warned you if it suspects you might be showing relapse behavior? For example, if someone was in recovery and did not want to relapse, an AI app could analyze their usage patterns/search history, sleep patterns, or other things, and then warn the person "you may experience greater than usual cravings or might relapse soon"? Not sure how practical or useful it would be, but it got me thinking.

Those are good feature ideas and some of them might not even require new data. Simply alteration of sleep patterns could issue warning, browsing new drug content sends major alarms. These are good for the MVP that gets seed funding. before the "magic" features can be implemented (like predictions from sufficient data)

 

[ownstrength]

I hope you know that gpt4 regularly hallucinates references and mechanisms. I use it for my writing sometimes and its something I'm incredibly cautious of

Yes but the migraine-anger-awareness is real, I did a quick perusal of articles, have a memory of reading about it, and it's much more often right than hallucinates. Although I am in many ways impressed by GPT4, it also made me stop worrying AI will take my job. Unless it plays dumb, but thats prolly projection. Started playing with the API to discover there are models with more context window up to 32K. I'm curious to create a long=term memory tnat loads stuff into the CW. I noticed convos with GPT4 were therapeutic and inspiring in various ways. Kind of like a skilled training buddy. A basic app I'd like to release for mental health is like an Emotional Buddy. Listens to your probs. talks to you helpfully, also creates a journak,etc. Starts learning to map your input to emotional state. It's a it of a grey area but it is often of similar effectveness as a mediocare counselor/therapist. Prefer not to make girlfriends... but if it makes money for other projects...