Perplexity "Deep Research" report on Cannabis health effects...

[old-fart]

I ran the following query on Perplexity's Deep Research. Here's the conclusion, but you can click the link to read the entire report.

based on current scientific knowledge, what are the positive or negative/adverse effects of smoking cannabis or ingesting THC edibles for the lungs, heart, vascular system, brain function and memory or any other health effects.

There is a lot more in the report beyond the conclusion, and I think it's an interesting read.

https://www.perplexity.ai/search/based-on-current-scientific-kn-AGmv3p_kR9ePJRlAcp4u4Q

Note: You don't need to sign in, just close the sign-in popup.

Conclusion​

Cannabis exerts multifaceted effects contingent on route, dose, and user age. Smoking imposes irreversible lung damage and cardiovascular stress, while edibles offer safer delivery for chronic conditions at the cost of dosing complexity. Neurocognitive risks—particularly in adolescents—warrant cautious prescribing despite potential neurological benefits. Future research must prioritize long-term outcomes of vaping, CBD/THC ratios, and receptor-specific agonists to optimize therapeutic indices. Policymakers should mandate standardized labeling and public education to mitigate preventable adverse events.

 

[old-fart]

Why do I care? I just got back my blood work for my annual physical, and there are few troubling areas that I will talk to my doctor about. I'm going down the internet rabbit hole to see if there is a connection between a few items in my results and cannabis use. So far, nothing concrete, but I'm learning good stuff.

 

[old-fart]

Another fact that may be of interest. I was curious how much tobacco is in a cigarette. Turns out most commercially manufactured cigarettes contain around 0.8 grams of tobacco. So when comparing tobacco and cannabis medical info, take this into consideration. I, for example, smoke about .1g/day. That would be similar to smoking 1/8 of a cigarette a day (I think), for purposes of lung damage. THC damage and holding the smoke in my lungs for longer than a tobacco smoker does are also factors.

From the research linked on the OP: "Regular cannabis smokers exhibit elevated rates of chronic bronchitis (OR 2.44 for morning sputum production), exercise-induced dyspnea, and wheezing. The Dunedin Multidisciplinary Study linked lifetime use to impaired gas transfer (DLCO reduction), indicating alveolar membrane damage. Notably, cessation reverses cough and sputum production but not hyperinflation or airflow resistance, implying irreversible structural changes"

hopefully (for me at least), the small amount I vape daily doesn't cause too much irreversible damage. Plus, I'm over 60, exercise regularly, so damage likely will take some time to build up. If you are young, you should consider the potential irreversible lung damage. The alternative, I guess, is edibles.

 

[buford]

Have you verified that the 20 papers cited by the robot both a) actually exist, and b) actually support the facts they are cited for?

If both those things are true, then you need to check the quality of the journals they were published in, and the reputations, and past work, and affiliations, and funding sources, of all the co-authors.

Then you might be in a position to judge whether the results this large language model predictive software has produced have anything at all to do with the real world, or have any value in considering issues around cannabis use and health. Probably not much value is my guess.

Even if all the papers exist, and actually stay what the robot says they say, i would urge skepticism in reading papers on a super-under-researched topic like this, with a huge structural/institutional bias towards finding harm (the US NIDA, the largest funder of cannabis research on the planet, may legally fund only studies intended to demonstrate harmful effects of use of illicit drugs, such as cannabis (on the Federal level)).

I would also apply the same skepticism towards any advice given to you about cannabis and health by most physicians. Definitely ask for specific citations to the papers they are basing their advice on. Then read the papers, and don't be surprised if your doctor is just as ignorant as the robot.

 

[old-fart]

I hane't yet read all the papers sited by the report, but they are sited and cross-linked in the report. Obviously one should never accept what an AI says as gospel, but it's a great start for learning more about any topic, including this one. I read a few of the papers, and the results are a bit concerning, but not enough to for me stop vaping. There is certainly enough information and links in the report that I will continue to follow the sited sources.

Did you read the report? I was pretty skeptical about AI/LLM answers. I'm pretty familiar with this stuff after 30 years as a senior IT architect at a National Lab, and my son is an AI professor at UPenn, so I ask him what I don't understand. The recently released "deep research" capability is quite impressive. I put together an anti-scam presentation for my local senior center and I used the deep research to propose a presentation, and it was pretty much spot on. I made a few changes, but I followed all the links, and had done a ton of research myself.

There is no doubt that LLMs and how to use them effetely is nuanced. I much more trust the results from a deep research query than I do from a cursory query, but I definitely get your point that the results should not be taken at face value.

I encourage folks to read the LLM output and follow links where a person might want more information or are concerned about the specific topic. It's not long report, and puts together the existing research nicely.

 

[chillAtGVC]

holding the smoke in my lungs for longer than a tobacco smoker

You may want to research that as well. My conclusion was that this is an "old wife's tale". Holding it in has no effect.

 

[buford]

No, i did not read the robot's output.

You mention that some of the papers caused you concern, but not enough to stop vaping. Does this imply that there was even one paper about health effects of vaping cannabis that was not about carts? I am not aware of much of any research at all being done on, for example, the health effects of dry-herb vaporizers. A quick check of PubMed confirms that there don't seem to be many/any papers.

This is not surprising for a route of administration that is very unpopular, but it does mean that there basically is practically zero guidance as to the potential health effects of dry herb vaporization to be gotten from the scientific literature at this point

 

[zeebudz]

I usually go right to pubmed to do med related searches but LLMs are getting better and are finally giving real references (they used to make up shit!)

Just asked deepseek your exact question. And check out the very beginning of the response (bolded mine). Note that vaping was unprompted in the question:

1. Lungs

Negative Effects:

Smoking Cannabis: Smoking cannabis, like smoking tobacco, can irritate the lungs and lead to chronic bronchitis symptoms (e.g., coughing, wheezing, and phlegm production). It may also cause inflammation and damage to the airways.

Carcinogens: Combustion of cannabis produces tar and carcinogens, which could increase the risk of lung cancer, though the evidence is less clear than with tobacco.

No Clear Benefit: Smoking cannabis does not appear to improve lung function and may harm it over time.

Positive Effects:

Alternative Methods: Using THC edibles or vaporizers (at lower temperatures) avoids lung irritation and may reduce respiratory risks compared to smoking.

 

[zeebudz]

BTW, I asked DeepSeek if its training included Granny Storm Crow's List, and it said no. But it did say it was trained in Pubmed to July 2024.

ChatGPT said it was aware of the list. I posed the same question and asked it to provide references. The references it provided were mostly from the Wiki article "Long-term effects of cannabis", web sites from the cdc and popular med news sites, and a single nih.gov article from the Journal of Drug and Alcohol Abuse, 2020. (The wiki article has references, but ChatGPT did not point to anything specific). For the respiratory system, it mentioned positive anti-inflammatory effects and edibles for avoiding respiratory harm. It also made no mention about vaping which DeepSeek did unprompted.

Here is the latest Granny Storm Crow's List:

Granny Storm Crow's List

If The Truth Won't Do, Then Something Is Wrong!

grannystormcrowslist.wordpress.com

 

[Flotsam]

some of these studies always seem suspect in their methodology. It seems vaping is never really addressed or if it is it talks about carts.

 

[TigoleBitties]

This is not surprising for a route of administration that is very unpopular,

Agreed here. Smoking involves burning things and science has long confirmed that inhaling smoke from things that are burning is bad for you.

but it does mean that there basically is practically zero guidance as to the potential health effects of dry herb vaporization to be gotten from the scientific literature at this point

There is very little in the scientific literature about DHV, that's true but I don't think the original query by @old-fart was well suited to come to any conclusions about the health effects of DHV as it specifically asked about smoking and edibles.

based on current scientific knowledge, what are the positive or negative/adverse effects of smoking cannabis or ingesting THC edibles for the lungs, heart, vascular system, brain function and memory or any other health effects.

Finding the best prompt or query to ask an AI is super important in determining the relevancy of the response. Here's what I asked Google Gemini off the top of my head.

"Tell me any health effects from low temperature vaping of cannabis in a vaporizer."

https://www.google.com/search?q=tel...QQIRgKmAMAkgcDMC4xoAeJEA&sclient=gws-wiz-serp

 

[old-fart]

Great discussion!

I would think that any study that covers THC in the system would still be relevant for dry herb vapors as DHV is just a delivery system. As for vaping vs smoking, you are still breathing in smoke, just less and possibly less harm smoke, but it’s still smoke.

 

[DeadRabb1t]

And there I was thinking that smoke and vapour are different things that occur via different processes

Great discussion!

I would think that any study that covers THC in the system would still be relevant for dry herb vapors as DHV is just a delivery system. As for vaping vs smoking, you are still breathing in smoke, just less and possibly less harm smoke, but it’s still smoke.

 

[justcametomind]

I think that AI nailed pretty well that there are no certain truths about vaporizing risks. It nailed it also about edibles as the only safe route. I think it's perhaps important to consider another issue with vaporizing and especially dabbing: it's the first time in history that is widely available a route that enables people to ingest so much cannabinoids via breath system all at once, and that opens the door to risks for the lungs, CHS and believe it or not (imho), addiction risks.

 

[buford]

There is very little in the scientific literature about DHV, that's true but I don't think the original query by @old-fart was well suited to come to any conclusions about the health effects of DHV as it specifically asked about smoking and edibles.

You are of course correct, and i didn't read the OPs prompt very carefully, but simply assumed that, if he was posting it here, that it would have some relevance to vaping, since this is a vaping forum. Of course, i was wrong and he was actually asking about non-vaping routes of administration, and then for some reason chose to post here.

My main point still stands: it was that asking Large Language Model text prediction bots for health advice is just incredibly stupid, and at best a waste of time. This is because they have a tendency to make shit up, and a tendency to misread sources, and have zero ability to distinguish high-quality work from hackwork. If you are inclined to generate LLM documents about health issues, be sure to check every source cited very carefully before relying on any purported health advice.

If you don't have the scientific background to search PubMed and understand the papers you get, you are thus wasting your time with an LLM. If you do have that background, you probably will have little inclination to use one. If you believe that LLMs can "answer questions", you do not understand how LLMs work.

 

[TigoleBitties]

My main point still stands: it was that asking Large Language Model text prediction bots for health advice is just incredibly stupid, and at best a waste of time. This is because they have a tendency to make shit up, and a tendency to misread sources, and have zero ability to distinguish high-quality work from hackwork. If you are inclined to generate LLM documents about health issues, be sure to check every source cited very carefully before relying on any purported health advice.

For sure, getting health advice from the internet is fraught with danger and even peer-reviewed research can be misleading and/or misinterpreted. Everyone needs to apply critical thinking and reading skills to any information received whether from an LLM or not. I think it's a bit extreme to say that asking LLMs for health advice is stupid. I agree with your last point though that now that most AIs offer a list of citations used in generating their answers, it can be useful to use the material to come to your own conclusions.

 

[florduh]

I don't have an opinion on AI analyzing health problems.

I'm more interested in giving AI health problems.

 

[old-fart]

Guys, stop talking about the efficacy of LLMs and focus on the content. The report is well cross referenced so if any of the items are concerning to you, as they were to me, use the report as a starting point for further research. This report didn’t make any anything up; it’s a pretty straight forward report. The idea is to comment on the results, now how I got them.

 

[florduh]

@old-fart IMO you'd get more mileage out of optimizing the other aspects of your health rather than focusing on cannabis consumption.

The study you posted yesterday mentioned both smokers and edible users experience a weakened endothelial system. How do you mitigate that? By doing things that improve overall cardiovascular health. Exercise, healthy eating, maintaining a healthy weight, and managing blood pressure. My guess is those things will also benefit the respiratory system.

 

[old-fart]

@old-fart IMO you'd get more mileage out of optimizing the other aspects of your health rather than focusing on cannabis consumption.

The study you posted yesterday mentioned both smokers and edible users experience a weakened endothelial system. How do you mitigate that? By doing things that improve overall cardiovascular health. Exercise, healthy eating, maintaining a healthy weight, and managing blood pressure. My guess is those things will also benefit the respiratory system.

Thanks. I agree that overall health, and the items you wrote about, is very important, especially cardio exercise. I already think about what I eat and I exercise daily. I really enjoy cannabis, and I prefer vaping to edibles. I (at my wife's prompting) want to learn more about the potential effects of vaping so I'm well informed. I'm a very strong believer in being well informed with actual facts, or at least the most up-to-date scientific data and theories.

As for mitigating a weakened endothelial system, well the report actually addresses that in the section "Mitigation Strategies". I don't know how efficacious the suggestion is, but it's certainly worth additional research. Has anyone heard of or use genistein? In the linked Stanford pretrial paper, weasel words like "may mitigate" are used. Then it goes on to say: "But genistein works quite well to mitigate marijuana-induced damage of the endothelial vessels without blocking the effects marijuana has on the central nervous system, and it could be a way for medical marijuana users to protect themselves from a cardiovascular standpoint." I find this interesting, and if taking a genistein supplement helps to mitigate some adverse affect of vaping, I want to learn more (and I will).

To be clear, I didn't post this to advocate not vaping cannabis. I ran the report to help educate myself on what I'm doing to my body, and I thought others would be interested. As for using Perplexity, I've been using it more and more for both simple and complex research and coding tasks, and I'm impressed with the results so far. It's a tool, nothing more.

Thanks.

 

[old-fart]

You may want to research that as well. My conclusion was that this is an "old wife's tale". Holding it in has no effect.

My understanding is the current thinking is about 2-3 seconds in your lungs is enough to absorb (right word?) most of the THC from the vapor. I asked the AI for a report :-)

Prompt: when vaping cannabis, how long should breath be held to maximize the effects of THC?

Conclusion
Current evidence conclusively demonstrates that breath-holding duration beyond 2–3 seconds provides no measurable increase in THC absorption during vaping. Optimal effects derive from controlled inhalation depth, device temperature management, and dosage titration rather than pulmonary retention time. Public health initiatives should prioritize correcting prevalent myths about breath-holding to reduce preventable cardiopulmonary risks in cannabis users. Future research directions include real-time pharmacokinetic monitoring using exhaled breath condensate analysis and development of inhalation-assisted devices with biofeedback capabilities.

I encourage people to read this report too. It's fascinating to me, and addresses vaping vs combustion (see the Bioavailability Differences Across Consumption Method section) and discusses "Vaping’s superior efficiency".

https://www.perplexity.ai/search/when-vaping-cannabis-how-long-EA9ppkq3R7O3b8CvItWE5Q

Again, just close the login windows if you don't want to log in. The report will still show.

And there I was thinking that smoke and vapour are different things that occur via different processes

Technically correct. Vapor and smoke are different things. My apologies. Still, inhaling cannabis vapor into the lungs is not natural for the lungs, and thus will have some effect on the lungs and cardio vascular system. I want to understand the non-phsycoactive effects to better maintain my health. Obviously there's more going on in my life, but understanding micro issues is as important as understanding macro issues.

 

[TigoleBitties]

To be clear, I didn't post this to advocate not vaping cannabis. I ran the report to help educate myself on what I'm doing to my body, and I thought others would be interested. As for using Perplexity, I've been using it more and more for both simple and complex research and coding tasks, and I'm impressed with the results so far. It's a tool, nothing more.

I think the fact that you're using AIs for to investigate things is great. I've been a developer for 30 years at a large Telco and for the last 2 years I've been doing R&D with AI models and frameworks. The pace of evolution is like nothing I've ever witnessed before in the tech space. If you think today's models are impressive, they will probably improve exponentially from here.

AI will no doubt dominate our future and I think we're at the inflection point right now. Learning how to interact with them effectively will be essential so I think all your early learning about them is valuable.

 

[old-fart]

I think the fact that you're using AIs for to investigate things is great. I've been a developer for 30 years at a large Telco and for the last 2 years I've been doing R&D with AI models and frameworks. The pace of evolution is like nothing I've ever witnessed before in the tech space. If you think today's models are impressive, they will probably improve exponentially from here.

AI will no doubt dominate our future and I think we're at the inflection point right now. Learning how to interact with them effectively will be essential so I think all your early learning about them is valuable.

I get an ad on Facebook for a University of New Mexico (state where I live) class on how to write a prompt ($1700)! I thought: how silly is that! But... I talked to some folks that are really into this and they tell me they use the AI to write their prompts, and their initial prompts for company financial info (for investing) or for coding a module are pretty long. It's a new world for information workers. Heck, UPenn even offers a Masters degree in AI (my son is not the professor this).

Most of my working life was spent as an application programmer, and then a solution architect, and then the chief IT architect at the Los Alamos National Lab (the one from the Oppenheimer movie). When I retired in 2023, AI availability was just getting going, and there's wasn't anything the Lab could use that met our security requirements. Just 2 years later, we have multiple on-site LLMs (we got some pretty good sized super-computers to run them on) as well as access to government versions of a few commercial LLMs. I've been told by folks that still work at the Lab these LLMs are changing both how science is done and how all knowledge workers get their work done.

Those are the talks that prompted me to use Perplexity to find out more about my new hobby.

 

[TigoleBitties]

I get an ad on Facebook for a University of New Mexico (state where I live) class on how to write a prompt ($1700)! I thought: how silly is that! But... I talked to some folks that are really into this and they tell me they use the AI to write their prompts, and their initial prompts for company financial info (for investing) or for coding a module are pretty long. It's a new world for information workers. Heck, UPenn even offers a Masters degree in AI (my son is not the professor this).

Most of my working life was spent as an application programmer, and then a solution architect, and then the chief IT architect at the Los Alamos National Lab (the one from the Oppenheimer movie). When I retired in 2023, AI availability was just getting going, and there's wasn't anything the Lab could use that met our security requirements. Just 2 years later, we have multiple on-site LLMs (we got some pretty good sized super-computers to run them on) as well as access to government versions of a few commercial LLMs. I've been told by folks that still work at the Lab these LLMs are changing both how science is done and how all knowledge workers get their work done.

Those are the talks that prompted me to use Perplexity to find out more about my new hobby.

Nice! I'm not yet retired but hope to be in about 5 years. I honestly can't imagine how pervasive AI will be by then given the breakneck pace things are at now. It doesn't surprise me that people use AIs to help them develop more effective prompts for the AIs themselves.

I found myself wondering how to design a specific AI agent recently and as I was sitting there thinking about it, I decided I'd ask a general AI about how to do it and yup, the results were useful.

 

[yogie]

Agreed here. Smoking involves burning things and science has long confirmed that inhaling smoke from things that are burning is bad for you.

There is very little in the scientific literature about DHV, that's true but I don't think the original query by @old-fart was well suited to come to any conclusions about the health effects of DHV as it specifically asked about smoking and edibles.

based on current scientific knowledge, what are the positive or negative/adverse effects of smoking cannabis or ingesting THC edibles for the lungs, heart, vascular system, brain function and memory or any other health effects.

Finding the best prompt or query to ask an AI is super important in determining the relevancy of the response. Here's what I asked Google Gemini off the top of my head.

"Tell me any health effects from low temperature vaping of cannabis in a vaporizer."

https://www.google.com/search?q=tel...QQIRgKmAMAkgcDMC4xoAeJEA&sclient=gws-wiz-serp

Thanks, this search helped me find this article, which supports vaping... https://pmc.ncbi.nlm.nih.gov/articles/PMC4456813/

I'm also interested or wondering in how the negative impacts compare to eating too much sugar, animal fat, etc.... I'm probably doing more harm to my body with sweets/sugar and bacon/fried chicken than dhv and tinctures.