Algorithms, addiction and anxiety: filters, ad blocking and why cutting off social media helps
Nathan Chadwick · · 14 min read
TLDRRead the short version
- Most feeds are built to maximise time on the app, not how you feel afterwards. Unpredictable rewards, endless scroll and personalised recommendations are very good at keeping people looking.
- The evidence is mixed on how big the harm is for the average person, but the experiments that took people off social media for a week or four found small, real improvements in mood and anxiety.
- Ad blockers and tracker blockers protect your privacy and your security. They do not remove the feed, so they are only half of the answer.
- Use layers: a good browser blocker, DNS filtering on your network, tighter phone settings, and filters that block the feeds themselves when you need them gone.
- Cutting off social media completely is easier than moderating it. Try a clean break for a week or two, then add back only what earns its place.
On this page
- The product is your attention
- What the algorithm is actually doing
- How it gets into us
- What the evidence does and does not say
- Why cutting off social media is useful
- How to block it: think in layers
- What ad blockers will not do
- If you want to try it: a simple plan
- If it is more than a habit
- The bottom line
- References
The product is your attention#
Most of the internet you use for free is paid for by advertising. That means the thing being sold is not the app. It is the time you spend looking at it, and what can be learned about you while you do.
Once you see it that way, the design of these products stops looking strange. Infinite scroll, autoplay, pull to refresh, streaks, notification badges and a feed that seems to know you a little too well are all doing the same job. They remove the natural stopping points, and they keep the next reward just out of sight.
I work in IT and security, so I tend to think about this as a systems problem. The system has a goal, and the goal is engagement. How you feel when you put the phone down is not one of the numbers it is tuned for.
I also know what it is like to be pulled into a system built to reward you. I wrote about that in how RuneScape was my escape. A game has an honest relationship with you, in that you know it is a game. A feed that is dressed up as your friends, the news and your hobbies is harder to see clearly.
What the algorithm is actually doing#
A recommendation algorithm is not trying to understand you. It is running an experiment on you, constantly. It shows you something, watches what you do, and shows you more of whatever made you stay.
What makes you stay is not always what is good for you. Some of the strongest signals are:
- Watching or re-watching. Pausing on something sad, upsetting or embarrassing counts the same as enjoying it.
- Outrage. A study of 12.7 million tweets and two experiments found that people who got more likes and shares for outrage went on to express more outrage, in line with reinforcement learning. The platform rewards it, so people produce more of it.
- Comparison. Content that makes you look at your body, your money or your life and come up short is sticky.
- Variable rewards. You never know whether the next swipe will be boring or brilliant. Researchers who study gambling have argued that this kind of unpredictability, along with infinite scroll and personalised recommendations, can give digital products addictive potential in the same way slot machines have.
None of that needs a villain. You only need a system that is rewarded for time on app and is very good at learning.
How it gets into us#
Addiction, or something like it#
Researchers still argue about whether "addiction" is the right word for social media use. It is not a formal diagnosis in the way gambling disorder is. But the pattern most people recognise is real: you open the app without deciding to, you stay longer than you meant to, you feel worse, and you do it again.
A one week field study of 46 people found that infinite scroll left people feeling caught in a loop that they regretted. Notably, most of the reasons people gave for stopping were outside the app, such as being interrupted or arriving somewhere. The app itself rarely gave them one.
Anxiety#
Anxiety rarely comes from one bad post. It builds up from:
- Constant comparison with the best moments of other people's lives.
- Doomscrolling, which keeps your threat detection switched on with no way to do anything about what you are seeing.
- Notifications that train you to stay alert for the next ping.
- Late night use that eats into sleep, which makes everything else harder.
In 2023 the US Surgeon General's advisory pointed to a study of adolescents that found those who spent more than three hours a day on social media faced double the risk of symptoms of depression and anxiety. The advisory is also honest that the evidence has gaps, and it says social media brings some benefits too.
Fixation#
This is the one that interests me most. A feed finds the thing you cannot look away from and then serves you more of it, faster than you could ever find it yourself. That can be a hobby, which is fine. It can also be a grievance, a body image worry, a conspiracy, an ex, or a mental health struggle.
Amnesty International and its technical partners tested this on TikTok. They set up automated accounts that simulated 13 year olds interested in mental health content. Within a short time, the feed was showing a heavy share of depressive and self-harm related videos. In their manual tests, more than half of the videos between 3 and 20 minutes in were about mental health struggles.
That is a test on one platform, with accounts built to show interest, so it does not tell you exactly what will happen to you. But it shows how quickly a feed can narrow.
My own view, and it is only my view rather than something the research proves, is that people whose brains already lock onto interests are probably more exposed. Many autistic and ADHD people describe intense focus as a strength. A system that is built to find your focus and then monopolise it is a different thing.
What the evidence does and does not say#
I do not want to oversell this, because the research is genuinely mixed.
- A large analysis of three datasets covering 355,358 adolescents found that digital technology use explained at most 0.4% of the variation in well-being, which the authors called too small to warrant policy change.
- Facebook's own researchers found that some teenage girls felt worse after using Instagram. Meta disputed how that was reported, noting that it was about girls who already had body image struggles and that the research was not designed to show cause.
- The Hunt et al. experiment found that limiting Facebook, Instagram and Snapchat to about 30 minutes a day reduced loneliness and depression over three weeks. A later replication did not find the same benefit, largely because most people did not stick to the limit.
So the honest summary is that for the average person the effect may be small, and for some people it is much bigger. If you already feel anxious, low, or unable to stop, you are more likely to be in the group that benefits from change.
Why cutting off social media is useful#
Look at the experiments where people were asked to stop, rather than just cut back.
- In a randomised trial of 154 adults, a one week break from Facebook, Instagram, Twitter and TikTok led to better well-being and lower depression and anxiety than carrying on as normal.
- In a four week Facebook deactivation experiment before the 2018 US midterms, people spent more time on offline activities, reported small but significant improvements in well-being, and kept using Facebook less after the experiment ended.
The effects were modest, not miraculous. The researchers on the Facebook study pointed out that the improvement was far smaller than a naive correlation would have suggested. But they are real, and they come from random assignment.
There is also a practical lesson hidden in the Hunt replication. When people were told to moderate, most could not. That matches how these products work. If every session is designed to have no natural end, then "just use it less" is a tough ask.
My take is that a clean break is often easier than moderation. You get to find out what you feel like without the feed, and you do not have to win a small fight with the app every hour.
Cutting off also gives you:
- Time back. The Facebook study found that deactivating freed up around an hour a day for the average person.
- A quieter head. You stop carrying the day's outrage and comparison around with you.
- Better sleep, if you let it. Not having a screen in bed is one of the easiest wins.
- A clearer view of what you actually miss. Some things you will miss. Many you will not. That tells you what to keep.
The trade-off is real too. The same Facebook study found that people who deactivated knew less about the news. Social media can also be how you keep in touch with friends and family. Cutting off does not have to be forever, and it does not have to be everything.
How to block it: think in layers#
No single tool does everything. Ad blocking, tracker blocking, DNS filtering and feed removal each deal with a different problem, so they work best together.
Layer 1: block ads and trackers in the browser#
This is the highest value change for most people.
- Use uBlock Origin on Firefox. The full version of uBlock Origin relies on browser features that Chrome has been phasing out. Firefox has said it is keeping the older extension model alongside the new one, so uBlock Origin keeps working there. If you stay on Chrome, the project offers uBlock Origin Lite, which is a different and more limited extension.
- Turn on tracking protection. Firefox has Enhanced Tracking Protection built in. Privacy Badger from the EFF is a good extra layer.
- Do it for security as well as privacy. In 2022 the FBI warned that criminals were buying search ads to impersonate brands and push malware, and it said to use an ad blocking extension when searching. If you have ever clicked the top search result for a download page, you will see why.
Layer 2: filter the whole network with DNS#
A DNS filter sits on your network and refuses to look up the addresses of known ad, tracker and malware domains. Because it works for every device that uses your network, it also covers smart TVs, games consoles and the apps on your phone.
- Pi-hole and AdGuard Home are free and run at home, on a Raspberry Pi, in a container or on a small server.
- NextDNS does the same thing as a hosted service, with no hardware, and it works on your phone when you are away from home.
- On a business or prosumer network, use the web or DNS filtering already built into your firewall. I covered the DNS side of this in DNS hygiene notes for a zero-trust network and the network side in UniFi segmentation.
DNS filtering has a hard limit. It can only block whole domains. Where ads come from the same domain as the content, it cannot tell them apart. The AdGuard Home project lists YouTube and Twitch ads and sponsored posts on Facebook, Twitter and Instagram as things a DNS blocker cannot remove. Pair DNS with a browser blocker for the rest.
Layer 3: tighten your phone#
Phones are where the tracking is hardest to see.
- iPhone: go to Settings, Privacy & Security, Tracking, and turn off Allow Apps to Request to Track.
- Android: open Settings, Privacy, Ads, and delete your advertising ID. The menu names change between versions, so search settings for "ads" if you cannot find it.
- Review permissions. Location, contacts, microphone and camera should be "only while using" or off for anything that does not need them.
- Use a private DNS provider on the phone such as NextDNS if you want the same filtering outside your home Wi-Fi.
These settings reduce cross-app tracking. They do not stop an app from learning from what you do inside it.
Layer 4: filters for yourself#
Most people think of internet filters as something for children. They work just as well when you are the one who needs protecting from yourself.
- Block the feeds on your network. AdGuard Home, NextDNS and most firewalls can block social networks by category or by domain, on a schedule if you like. Block them at night, or entirely for two weeks.
- Block them on your computer. LeechBlock NG is a free browser extension that lets you set time limits and schedules for sites.
- Make it hard to undo. A filter you can switch off in five seconds will be switched off in five seconds. Ask someone you trust to set the filter password and keep it for the length of your break.
- If you have children, UK broadband providers offer free network level parental controls, and Internet Matters has step by step guides. Since July 2025 the UK's Online Safety Act has also required services likely to be used by children, where they pose a medium or high risk, to configure their algorithms to filter harmful content out of children's feeds. That helps, but it does not replace a conversation at home.
Layer 5: take the algorithm out of the feed#
This is the layer that targets the actual problem, so it is the one that matters most.
- Turn off autoplay wherever you can, and pause your watch history on video sites. Both take away a lot of what drives recommendations.
- Switch to a feed that is not personalised. Many apps offer a "following" or chronological view. In the EU, the Digital Services Act requires the biggest platforms to offer at least one feed option that is not based on profiling. That law covers the EU, so what you see in the UK may differ.
- Remove the apps from your phone and use the website in a browser instead. It is slower and has fewer hooks, which is the point.
- Turn off every non-human notification. Messages from people can stay. Everything else can go.
- Set the screen to greyscale in your accessibility settings. It takes some of the shine off.
- Charge the phone outside the bedroom. Buy an alarm clock.
- Log out after every session. The extra friction is the whole idea.
What ad blockers will not do#
It is worth being clear about this, because it is easy to think a blocker fixes the problem.
- Blocking ads does not remove the feed. Ads are one way these platforms make money. The algorithm is the product.
- Blocking trackers does not stop on-platform profiling. If you are logged in, the app can see everything you do inside it.
- Blockers do not reduce the pull. You can have a clean, ad free, tracker free feed and still lose three hours to it.
Blocking is worth doing for privacy and security alone. But if what you want is to feel better, the feed itself has to change, or go.
If you want to try it: a simple plan#
- Pick a window. One week is enough to notice a difference, and two is better. Tell a friend so that someone else knows.
- Set up the browser and DNS layers first. They take an hour and they are worth keeping for good.
- Block the feeds you lose the most time to. Use a filter that is hard to switch off.
- Remove the apps from your phone, and turn off notifications from everything that is not a person.
- Replace the habit. Decide what you will do in the gaps. A walk, a book, a podcast, the gym, a call. If the gap is empty, the phone will fill it.
- Expect it to feel odd. You will reach for the phone without thinking. That is the habit showing itself, not a sign that you are failing.
- At the end, add back only what earns its place. Keep the group chat. Keep the creator you actually enjoy. Leave the rest.
If it is more than a habit#
If you have been struggling with anxiety, low mood or compulsive use for a while, it is worth talking to someone. A GP, a therapist or a friend can all help. The tools above can make life easier, but they are not a replacement for support.
If you are struggling right now: Samaritans on 116 123, any time, free. NHS 111 and pick the mental health option. Mind has information and an infoline. If you are in danger, call 999.
The bottom line#
You are not weak for finding this hard. A great deal of money and a great deal of machine learning has been spent on making it hard.
You can make it easier by putting things between you and the feed: a blocker in the browser, a filter on the network, a few changes on the phone, and, when you need it, a hard stop. You do not need to be perfect. You need to put enough friction in the way that the choice is yours again.
References#
- Allcott et al., The Welfare Effects of Social Media (American Economic Review, 2020)
- Lambert et al., Taking a One-Week Break from Social Media Improves Well-Being, Depression, and Anxiety (2022)
- Hunt et al., No More FOMO: Limiting Social Media Decreases Loneliness and Depression (2018)
- Replication of Hunt et al.: Is Limiting Social Media Use Feasible for Young Adults? (preprint)
- Orben and Przybylski, The association between adolescent well-being and digital technology use (Nature Human Behaviour, 2019)
- US Surgeon General, Social Media and Youth Mental Health advisory (2023)
- Brady et al., How social learning amplifies moral outrage expression in online social networks (Science Advances, 2021)
- Amnesty International, TikTok's For You feed risks pushing children towards harmful mental health content (2023)
- Meta, What our research really says about teen well-being and Instagram (2021)
- Engineered highs: reward variability and frequency as potential prerequisites of behavioural addiction (Addictive Behaviors, 2023)
- The Loop and Reasons to Break It: Investigating Infinite Scrolling Behaviour (Proceedings of the ACM on Human-Computer Interaction, 2023)
- FBI IC3, Cyber criminals impersonating brands using search engine advertisement services (2022)
- BleepingComputer, Firefox continues Manifest V2 support as Chrome disables MV2 ad-blockers
- uBlock Origin Lite FAQ
- Mozilla, Enhanced Tracking Protection in Firefox for desktop
- EFF, Privacy Badger
- AdGuard Home, known limitations of DNS-level blocking
- Pi-hole documentation
- Apple Support, If an app asks to track your activity
- EFF, How to disable ad ID tracking on iOS and Android
- Ofcom, New rules for a safer generation of children online
- Digital Services Act, Article 38: non-profiling recommender options
- Internet Matters, Parental controls
- LeechBlock NG
- NextDNS
- Samaritans
- Mind