How to Overcome Addiction to Substances or Behaviors | Dr. Keith Humphreys
This comprehensive discussion between Dr. Andrew Huberman and Stanford psychiatrist Dr. Keith Humphreys covers the science of addiction across substances (alcohol, cannabis, opioids, nicotine) and behaviors (gambling, social media), exploring genetic risk factors, the truth about health claims, and evidence-based recovery strategies including 12-step programs, emerging treatments like psychedelics and GLP-1 medications, and practical tools for talking to someone struggling with addiction.
1. Understanding Addiction: More Than Just a Bad Habit
Dr. Keith Humphreys, a professor of psychiatry and behavioral sciences at Stanford School of Medicine, brings decades of experience treating addictions and shaping drug policy. He begins by clarifying what addiction actually means versus how we casually use the term.
"It's not just stuff you do a lot, which we sometimes colloquially call addiction. It's the persistence of doing something that is harmful."
He references classic animal studies where rats would press a button for brain stimulation even while starving to death next to food pellets. That's the essence of addiction—continuing a behavior to the point of destruction when anyone would normally just stop.
Dr. Huberman offers his own working definition:
"It's a progressive narrowing of the things that bring one pleasure."
Dr. Humphreys agrees, explaining how addicted individuals sacrifice relationships, work, and housing for their substance. Eventually, the drug becomes the only rewarding thing left in their life, making it even harder to quit.
2. The Genetic Factor: Are You Born at Risk?
A major question addressed is whether addiction is genetic. Dr. Humphreys first dispels a common myth:
"We say people are born addicted. You'll sometimes read if mom was addicted to fentanyl, then the baby is born addicted. That is not possible."
A fetus can be physically dependent (experiencing withdrawal at birth) but not addicted, since addiction requires a learned association between behavior and drug effects. However, genetic risk from birth is very real and substantial—studies of adopted children show they're much more likely to develop alcohol problems if their biological parents were alcoholics, even when raised by non-drinkers.
Some genetic factors are substance-specific. For example, people of Han Chinese descent who lack certain alcohol-metabolizing enzymes find drinking unpleasant, lowering their risk. Other genes affecting impulsivity or sensation-seeking create broader vulnerability across multiple substances.
The Subjective Experience Varies Wildly
Dr. Humphreys shares a personal example:
"I find taking opioids so unpleasant. I feel bound up, miserable, groggy that I just took one and said, 'Pain is better than this.'"
Yet he's worked with patients who describe their first opioid experience completely differently:
"The first time I had an opioid, it was like a hole in my chest that had been there my whole life filled up for the very first time."
This difference is largely genetic—and it explains why some people can take a substance once and walk away while others are immediately hooked.
3. Alcohol: Separating Myth from Reality
The "Red Wine Is Good for You" Myth
Dr. Humphreys directly addresses the persistent claim that moderate drinking, especially red wine, provides health benefits:
"Statement against interest because I like red wine. I would love to believe it is healthy. It's not."
The famous J-shaped curve showing moderate drinkers living longer than non-drinkers has a major flaw: non-drinkers include people in recovery from alcoholism who already have health damage. They're not dying because they stopped drinking—they had to stop because of prior harm.
The key insight Dr. Humphreys offers:
"There might be some cardiac benefit, okay? But we don't get to live our lives as single organs. We have a whole body. You have to weigh that if that is true. And it is wobbly. If that's true, it's smaller than the cancer risk."
Bottom line: Two drinks per week carries very small additional risk, but zero is definitively better than any amount for overall health.
Women and Alcohol: An Industry Strategy
Dr. Humphreys reveals a troubling trend:
"In the late '90s, early odds, the alcohol industry figured out that women had more money and but they weren't drinking the way men were. So they engaged in a long-term campaign to try to increase women's drinking."
The "mommy wine culture" on social media was often engineered by the industry. The result? Women's drinking increased significantly, despite women experiencing greater harm per drink than men. Encouragingly, younger generations are now questioning this trend.
The Business Model of Addiction
A startling statistic: 10% of Americans drink about half of all alcohol consumed. This means the alcohol industry's profits depend heavily on problem drinkers.
"If you're running the industry, you want that group to be as big as possible. You do not make money off people who have a half a bottle of wine on special occasions."
4. Cannabis: Not Your Parents' Pot 🌿
Dr. Humphreys draws a sharp distinction between "old" and "new" cannabis:
- 1980s-90s: Average THC content was 3-5%
- Today: Average product is about 20% THC
- Usage patterns: 42% of current users consume daily or almost daily
When you combine higher potency with more frequent use, brain exposure is dramatically different.
"The potency difference between a coca leaf and cocaine is 65 times. That is also the difference between old and new cannabis consumption patterns."
Why Parents Need to Update Their Thinking
"It's very hard to get across to parents because their view is like, 'Ah, I smoked weed, who cares if my 15-year-old is using it.' It's like but that's kind of like saying you drank low alcohol beer and you're not concerned that your 15-year-old is guzzling vodka."
The Psychosis Risk
Dr. Humphreys admits he was skeptical of the cannabis-psychosis link for years but has changed his view:
"The evidence has gotten stronger as the drug has gotten stronger... I would not use cannabis if I had any first-degree relatives with any schizophrenia, schizoid personality, anything in the psychotic spectrum, bipolar disorder."
Edibles Present Special Risks
While Dr. Humphreys disagrees that people are good at gauging their high from smoked cannabis, he acknowledges edibles create unique problems:
"When you eat something, it takes a while to have its effect. Particularly when these products came out, people would bite down on one piece, five minutes later feel the same, take another bite... and then it would all hit them like a train."
5. The Gateway Drug Question and Decriminalization vs. Legalization
All Drugs Are Gateway Drugs
"The lie in that was that cannabis had some unique role. But the truth is anything—if you're a teenager and you start smoking or you start drinking or you start using cannabis—that will increase your likelihood of progressing to other substances."
The mechanisms include: enjoying intoxication and wanting to try more, changing social networks to include other users, and possible brain sensitization.
The Alcohol Hypocrisy
"Alcohol is a drug and we pretend that it isn't. I have seen conferences, political events where people spend all day demonizing drug users... and then they all go to the bar and get drunk as if they are not drug users."
Decriminalization vs. Legalization: A Critical Distinction
- Decriminalization: Removes penalties for users but doesn't create a legal industry
- Legalization: Makes production, marketing, and sale legal, bringing in corporations
"The corporation is going to have very smart people who are good at selling and they will increase consumption of the product."
6. Gambling: The Hidden Epidemic 🎰
Dr. Humphreys expresses alarm at how gambling restrictions have essentially disappeared:
"When I was a kid, Pete Rose was not allowed to go into the Hall of Fame because he had once placed a bet on his own team. Now, you can't watch a sporting event without having gambling ads shoved in your face."
The Science of Slot Machine Addiction
Drawing from the book Addiction by Design, Dr. Humphreys explains how modern slot machines work:
"Dealers don't give the perfect timing of reinforcement that machines can do... A machine can give you exact timing between your press the button and then you get your reward."
The industry discovered that people aren't actually playing to win—they're playing for novelty. Modern slot machines use infinite combinations of symbols to create constant novelty, keeping people playing even while they're objectively losing.
"Losses disguised as wins. You put in a dollar and get a hundred credits, then it goes off—you've won 40, 20, and 10 with all these exciting things. I just lost 30% of what I put in. But it feels like a win."
7. Psychedelics, Ketamine, and Ibogaine: Emerging Treatments
Psilocybin and LSD
Dr. Humphreys sees genuine promise in psychedelic research:
"They're exciting in part because we haven't really made much progress in pharmacotherapy in the last 20 years for lots of things—for depression, for addiction."
Key points about the research:
- Small studies are encouraging
- Thankfully, psilocybin and LSD have extremely low addiction potential
- Microdosing has zero evidence of benefit—only high-dose clinical protocols show promise
- Proper trials require preparation sessions, guided experiences, and integration afterward
Why Most Research Happens with Psilocybin, Not LSD
A practical reason: LSD trips can last up to 13 hours, making staffing clinical trials extremely difficult. Most LSD research happens in Switzerland, where researchers work longer hours.
Ketamine: Promise and Problems
While FDA-approved for treatment-resistant depression, Dr. Humphreys notes:
"There are a lot of negative trials for depression. It didn't vault over the efficacy thing. It cleared it."
The downsides are significant: ketamine is addictive and causes bladder damage. Many young people now have "60-year-old bladders" from ketamine use.
The SAINT Protocol: A Better Alternative
Dr. Humphreys highlights the work of their late colleague Dr. Nolan Williams on repetitive transcranial magnetic stimulation (rTMS):
"The effects of that for treatment-resistant depression are so much clearer in my view, and the downsides are, as far as I can see, virtually nil."
Unlike psychedelic trials, rTMS studies can truly be blinded—participants can't tell if they're receiving the real treatment, making the science more rigorous.
Ibogaine: A Promising but Risky Frontier
Dr. Williams was pioneering ibogaine research for veterans with PTSD and addiction before his untimely death. This 22-hour psychedelic experience requires intensive cardiac monitoring and shows remarkable results in open-label trials, though controlled studies are still needed.
8. Stimulants and Nicotine: The Disappointing Treatment Landscape
Stimulant Addiction: 40 Years of Little Progress
Dr. Humphreys calls this his biggest career disappointment:
"The treatment offering to people who were addicted to crack cocaine then in the late '80s is not very different from what it is today, almost 40 years later. No pharmacotherapy at all."
The only thing that consistently works is contingency management—providing immediate rewards for clean drug tests. This proves that addicted people do have control; it's just impaired, not absent.
Prescription Stimulants: A Complicated Picture
"There are kids whose lives are transformed positively by Ritalin... At the same time, I would say overprescribed."
Dr. Humphreys worries about decreased tolerance for natural variation in how brains work:
"A kid can't be shy anymore. They have to be on the spectrum... and carry a diagnostic label."
Nicotine: Mistaking Withdrawal for Benefit
Many people think nicotine helps them focus and relax simultaneously. Dr. Humphreys offers a different interpretation:
"A lot of people who use it are mistaking the treatment of withdrawal for a drug benefit."
When you wake up after not smoking all night, you feel jittery. That first cigarette feels great—but it's just relieving withdrawal, not providing actual enhancement. This same pattern occurs with cannabis (sleep) and opioids (pain).
9. How to Talk to Someone with Addiction 💬
First Steps: Create Safety and Hope
When someone admits they might have a problem, Dr. Humphreys recommends:
-
Validate their courage: "Wow, I'm so glad you told me. This is something that tens of millions of people experience."
-
Convey optimism: "There are probably 24 million Americans are in recovery. We just don't notice them because someone in recovery looks like anybody else."
-
Emphasize early intervention: The odds of recovery are dramatically higher when people still have family support, housing, and employment intact.
The Surprising Question: "Why Would You Want to Quit?"
"Someone says, 'I want to quit smoking.' A good clinician will say, 'Why? Why would you want to do that?'"
This seems counterintuitive, but it's crucial. If the motivation doesn't come from within, it won't stick. Help them articulate their own reasons:
"Tell me why would you want—what do you want to get out of this? Because it's work."
Help them build vivid, personal motivations: "So if you had $2,000 because you hadn't smoked in a year, what would you buy for yourself? Tell me about it."
Conduct a Behavioral Analysis
Ask questions like:
- Where and when do you use?
- What triggers use?
- Are there places you would never use? (Mom's house? Holy days?)
- What techniques help you get through non-use situations?
"Could that go somewhere else? Could you give that away so that it's behaviorally harder for you to get this?"
The Power of Community
"Hang out with other people who are trying to make the same change. You want to start jogging? Join a jogging group. You want to stop drinking? I would suggest go check into an AA meeting."
This provides both support and accountability:
"Hey, you were going jogging Tuesday, you weren't there. What's up? Are you going to be part of this group or not?"
10. Understanding Addicts and Those Who Love Them
It's Not a Character Defect
Addiction was historically viewed as moral weakness. Dr. Humphreys explains why this view persists:
"A lot of the explanations for addiction come from people who are hurt and angry with good reason. They had an addicted parent and that was hard for them, or their marriage is disintegrating."
Addicted people do harm others—lying about money, missing important events, even putting children at risk. But when they get sober:
"It's spectacular how the real person seems to emerge, which points to the fact that the addiction masks something about who they truly are, not the other way around."
The "Codependent" Label May Be Unfair
Research by Ruth Kronhite showed that women married to alcoholic men displayed many "codependent" behaviors—but when their husbands got sober, they looked exactly like women whose husbands had never been alcoholic.
"A lot of things that are attributed to the personality of the codependent person is actually reaction to addiction... They're hyper-responsible. They have to be because the mortgage won't get paid."
11. Why Immediate Rewards Matter More Than Future Benefits
All people discount future rewards to some extent—we buy the $5 latte instead of saving for retirement. Addicted individuals do this even more extremely.
"In addiction, if you ask people, 'Would you take $5 today or $20 tomorrow?' they're more likely to say '$5 right now.' Almost as if tomorrow doesn't exist."
This is why long-term goals—marriage, career, health—aren't enough to motivate change. Focus on immediate benefits:
- More money every day
- Zero risk of arrest (for illegal drugs)
- Feeling physically better right away
- Social recognition for days/years sober
The genius of 12-step programs is the one day at a time concept:
"You can't suddenly quit drinking for the rest of your life. It's not here yet, right? And that just seems inconceivable. But can you not drink today? Not drink today and go to a meeting and get some reward for that? Yeah, you can probably do that."
12. Brain Changes: The Cue-Elicited Craving Problem
Long-term drug use physically changes the brain. Dr. Humphreys and colleagues conducted brain imaging studies on people in residential treatment for methamphetamine addiction:
"Showing them cues of meth-associated things—like the pipe or the powder—and asking them, 'How much do you like that?' Well, independent of that, there's also nucleus accumbens activation that you can see. And that predicted who relapsed. Not what they said, but what was going on in their brain."
This helps explain why addicted people sometimes seem to lie about their desire to quit:
Need a summary like this?
Get instant summaries with Harvest