International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 13 Issue: 03 | Mar 2026 www.irjet.net p-ISSN: 2395-0072
Role of Artificial Intelligence in Improving Mental Health Care
Shruti
Kapure1, Gausiya Shaikh2, Divya Deshmukh3, Vaishnavi Ankade4
1,2,3,4
Final Year Student, Dept. of Computer Engineering, P.K. Technical Campus, Maharashtra, India Guide: Prof. Tejashri Patil, HOD, Dept. of Computer Engineering
Abstract Mental health issues are increasing at a pace that our existing healthcare systems were never designed to accommodate. There's a huge number of people everywhere in the world, and largely here in India, who need psychological support of some kind but are unable to access it either due to the lack of trained professionals or for reasons of cost, location or stigma. This paper explores one potential solution to that gap: artificial intelligence. The study looks at published studies and a small primary survey of 50 college students from Maharashtra to gauge levels of awareness, usage and concerns young people have about AI-based mental health tools. The reality is that AI-based tools are genuinely promising chatbots, mood tracking apps and early detection systems are already helping people who might otherwise go unsupported. But there are serious issues as well, especially relating to privacy, the danger of erroneous answers in sensitive situations and the simple fact that a chatbot can never truly replace another human. This paper tries to look at both sides without overselling the technology.
Key Words: Artificial Intelligence, Mental Health Care, AI Chatbots, Mood Tracking Application, Early Detection of Mental Health Issues, AI-based Mental Health Tools, Privacy Concerns in AI
1. INTRODUCTION
Thinkaboutthetimeyoufeltsuperstressed.Maybebeforeabigtestorwhenthingsweretoughathome.Insituationslike that most people don't think about calling a therapist. They usually talk to a friend search online for answers or just sit quietly and hope the stress goes away. It's not that therapy doesn't work. The real issue is that for people in India and other developing countries getting professional help, for mental health is not easy. There aren't psychiatrists to support everyonewhoneedshelp.Thecostcanbehigh.Tobehonestpeoplestillfeeluncomfortabletalkingaboutmentalhealth.
Thatisinwhichera,artificialintelligencehasstartedouttoplayafunction.ArtificialintelligenceorAIislargelysoftware thatlearnsfromdata,findspatternsandmakesdecisionsorsuggestions.WealreadyuseAIeachdaywithoutknowingit. Forinstance,whenamusicappsuggestsatrack,orwhenabankstopsatransactionorwhenyoursmartphonecorrectsa spellingmistake.Thesametechnologyisnowbeingusedtobuildtoolsthatcanlistentosomeonedescribetheiremotions andreplyinameaningfulway.
Mentalhealthisaproblem.TheWorldHealthOrganizationsaidin2022thataboutoneineightpeoplearoundtheworld havesomekindofmentalhealthissueandmostofthemdonotgetanyhelp.InIndiathenumbersfromNIMHANSin2021 show that around 150 million people need mental health care but there are fewer than 9,000 psychiatrists in the whole country.Thisisaproblem.Thenumbersjustdonotaddup,anditwillnotgetbettersoonifweonlydothingsthewaywe alwayshave.ArtificialIntelligenceisnotasolutiontothisproblem,butitcanbeausefultoolifweuseitinacareful and responsiblewayanditcouldreallyhelpmakeadifferenceinmentalhealthcare.[1][7]
This paper explores how AI is currently being used in mental health care, what the research says about it, what young peopleintechnicaleducationthinkaboutit,andwherethereallimitationslie.Thegoalisnottohypethetechnologybutto giveagrounded,honestassessmentofwherethingsstandrightnow.
2. LITERATURE REVIEW
Peoplehavebeentryingtoautomatesupportforalongtime.Inthe1960sacomputerscientistJosephWeizenbaummade aprogrammecalledELIZAatMIT.Itwasachatbotthatwouldrephrasewhatuserssaidbacktothemasaquestion.This waslikewhatatherapistmightdo.ManypeopleweresurprisedwhentheyfoundoutthatELIZAwasaprogramme.What surprisedWeizenbaumwasthatpeopleformed emotional bondswithELIZA.Hetoldthemitwasa programme still they gotattached.Thisreactionhadanimpactonresearchersformanydecades.
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 13 Issue: 03 | Mar 2026 www.irjet.net p-ISSN: 2395-0072
The real growth in the health support area came with smartphones. Smartphones made it possible for people to have computersintheirpockets. Thismeant thatpeoplecouldgethealthsupport whentheyneededit.One study thatpeople talk about a lot is the work that Fitzpatrick and colleagues did in 2017. They worked with college students who had symptoms of depression and anxiety. These students used a chatbot called Woebot. The study was set up so that some studentsusedWoebotandothersgotalistofthingstoreadinstead.AftertwoweeksthestudentswhousedWoebothad lowerdepressionscores.The people whodidthestudy werecareful.Theysaidthat thegroupofstudentswassmall and they only did the study for a short time. The study was still important. It showed that the idea of using a chatbot like Woebotforhealthsupportwasworthconsidering.[2]
Wysaisanotherappworthlookingat.ItwasdevelopedinIndiaandtestedonpeoplethere.Astudyin2018byInksterand colleaguescheckedonWysa usersovertime.TheyfoundthatpeoplewhousedWysaoftenhadfeweranxietysymptoms. Wysa uses a mix of ways to help people. It has behavioural therapy techniques, mindfulness exercises and mood journaling. The NHS in the UK now uses Wysa as a digital resource. This shows that health systems are starting to take thesetoolsseriously.[3]
ResearchersarealsolookingintowhetherArtificialIntelligencecanfigureoutifpeoplehavehealthissuesbythewaythey talkorwriteontheinternet.Forexample,astudybyCoppersmithandcolleaguesin2018foundthatlanguagepatternsin social media posts can be used to identify signs of suicidal thoughts with pretty good accuracy. This kind of research is promising.Itisalsouncomfortable.Itmakesusthinkaboutissueslikeconsentandsurveillance.[4]
Intheyear2020ateamofresearchersledbyLinardondidastudy.Theylookedat72healthappsthatyoucandownload from app stores. What they found out is that very few of these apps less than 4 percent were tested to see if they reallywork.Soeventhoughtherearehundredsofhealthappsthatsaytheycanhelpyou,almostnoneofthemhavebeen proventobeeffective.[5]
Chancellor and De Choudhury also pointed out in 2020 that many artificial intelligence models used for health were trainedusingdatafromWesterncountriesandpeoplewhospeakEnglish.Thismeansthatthesemodelsdonotworkwell forpeoplefromotherculturesandlanguages.ThisisanissueforacountrylikeIndia,wherementalhealthappsareused bypeoplewhospeakmanydifferentlanguagesandcomefromdifferentbackgrounds.[6]
3. METHODOLOGY
3.1
Type of Research
Thisstudyisdoingthingsalittledifferently.Itisusingamixedapproach.Thepartwherewelookedatwhatotherpeople havefoundoutinvolvedreadinga lotofpublishedpapers,reportsfromthe World HealthOrganizationandarticlesfrom Google Scholar. We mostly looked at things that were written between 2015 and 2025. The other part of the study involvedaskingstudentswhattheyreallythinkaboutAIhealthtools.Wewantedtoknowwhatactualstudentsthink,not justwhatthepapersandarticlessay.
3.2 How the Survey Was Done
AGoogleFormquestionnairewasgiventofinalyeardiplomaandengineeringstudentsinthePimpri-ChinchwadandPune area in February 2026. This Google Form questionnaire was sent to the students through college WhatsApp groups and email.TheGoogleFormquestionnairedidnotaskfornamessostudentscouldanswerthequestionshonestly.Wegot54 responses.Outofthese54responses,50werecompleteandcouldbeused.Theother4responseswerenot complete,so theywerenotused.
The participants were aged between 18 and 25. All had a technical education background. We chose this group because theyarelikelytobeamongthefirsttotryouttechnology.Theyarealsoatanagewheretheyoftenfacealotofstressand anxietyabouttheirstudiesandothermentalhealthissues.Thesurveyhad15questions.Wefocusedonfivekeyonesfor ouranalysis.
3.3
Sample Questions from the Survey
Q1.Iwanttoknowifyouarefamiliarwithappsorartificialintelligencetoolsthat aredesignedtohelppeoplewithhealth issues.Doyouknowaboutthesetools?(Yes/No/Somewhat)
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 13 Issue: 03 | Mar 2026 www.irjet.net p-ISSN: 2395-0072
Q2.HaveyoueveractuallyusedanapplikeWysaorWoebotorHeadspaceorsomethinglikethese?(Yes/No/Heardof thembutnevertried)
Q3. If you were going through a tough time, would you be comfortable talking to an artificial intelligence chatbot about whatyouaregoingthrough?(Yes/Yesbutwithsomedoubts/No/Notsure)
Q4.HowworriedareyouaboutyourpersonalinformationbeingmisusedifyouuseanAImentalhealthapp?(Notworried /Alittleworried/Moderatelyworried/Veryworried)
Q5.Ifyouarefeelingalittlestressedandyouhavetwooptions anAIchatbotorahumantherapist andbotharefree andavailablerightnow.Whichwouldyouprefer?(AIchatbot/Humantherapist/Nostrongpreference)
4. APPLICATIONS OF AI IN MENTAL HEALTH
4.1 AI Chatbots
ThethingthatpeopleknowthemostaboutArtificialIntelligenceinhealthisthechatbot.Youcantalktotheseapps,type to them and they talk back to you like a person. The good ones, like Woebot, Wysa, Replika do not just say anything. ArtificialIntelligencechatbotsareusingtechniquesthattherapistsuse,likeCognitiveBehaviouralTherapy.Theyaskyou questionsinacertainorder.Theyalsohelppeopledoexercisesthattherapistsdowithpatients.Theydoitinagentleway. [2][3]
Thebiggestthingthatchatbotsofferisthattheyareavailable.Someonewhoishavingapanicattackat2AMcannotcalla psychiatrist.Thispersoncanopenanappontheirphone.Thereisnowaitingandnoneedtomakeanappointment.This persondoesnothavetodealwithafirstsessionwitha stranger.Forpeople,especiallyyoungpeoplewhogrewupusing screens,usingachatbotisamorenaturalfirststepthanwalkingintoaclinic.Thethingthatworriespeopleaboutchatbots isthattheysometimesmissthings.Chatbotscanmisreadasituation,givearesponsethatdoesnotfitorfailtopickupon somethingserious.Chatbotsareastartingpointandnotacompletesolutionforpeoplewhoneedhelp.
4.2 Mood Tracking
Thereareappsthataskyoutotellthemhowyouarefeelingeveryday.Youjustmusttaponascalethatsayssadorhappy. Someappsaremoreadvanced.Theylookatthingslikehowyouuseyourphoneandwhattimeyougotosleep.Theyeven checkhowmuchyouaremovingaroundandhowyousoundwhenyoutalktothem.Thesethingshelptheappsseewhat youareusuallylike.Ifsomethingchanges likeyouarenotmovingaroundmuchoryouaresleepingatweirdtimes the app will notice. It can even tell if you are not talking to people as much as youusually do. The app can then say that somethingmightbewrongwithyou.Theappsusethesepatternstofigureoutifyouneedhelpornot.
Thiskindofwarningishelpfulforthingslikebipolardisorder.Whenyoucatchachangeinyourmoodearly,itcanstopa bigger problem from happening later. For people who use this a lot it is a way to keep track of how they're feeling over time. You can see how you are doing before things get bad, instead of only noticing when bipolar disorder is really affectingyou.
4.3 Early Detection
Some AI researchers are working on spotting health issues early. They train machine learning models to find signs of problemsbeforepeoplenoticethem.Onewaytheydothis isbyanalysingspeechrecordings.Theylookathowsomeone talks.Theycheckifthetoneofvoiceisflat.Theyseehowvariedthewordssomeoneuses.Allthesethingscanbesignsof depression. Researchers also study social media posts. They search for patterns in what people write and how people write.Thishelpsthemfindmentalhealthissuesearly.Thegoalistohelppeoplegetsupportsooner. [4][5]
Ina real-lifesituationthiscouldmeanthata collegecounsellorusesanAItoolatthebeginningofa semester tofindout whichstudentsmightneedsomehelp.Adoctormightalsouseatoollikethiswhentheyaredoingacheckuponsomeone. Theartificial intelligencetechnologyisnotgoodenoughtosayforsure ifsomeone hasa problemonitsown. Itcanhelp peoplefigureoutwhoneedstobelookedatmoreclosely.Thisiswhattheartificialintelligencetechnologyisgoodfor it helpspeopledecidewhoneedsacloserlook.
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 13 Issue: 03 | Mar 2026 www.irjet.net p-ISSN: 2395-0072
4.4 Virtual Therapists
Virtual therapist systems are a step beyond chatbots. These are animated characters that can have voice conversations withpeople.Theycaneven respondtofacial expressions.Forexample,thereisELLIE,whichwasbuiltby researchersat theUniversityofSouthernCalifornia.TheresearcherstestedELLIE withveteranswhoaredealingwithPTSD. Whatthey foundoutwassurprising.TheseveteranstoldELLIEmoreabouttheirsymptomsthantheytoldhumaninterviewers.The reasonforthisseemstobethatpeopledonotfeeljudgedbyanon-human,sotheyaremorehonest.Thismightsoundabit weird,butitmakessenseinaway.Manypeoplefinditeasiertowritedowntheirthoughtsandfeelingsinadiarythanto saythesamethingsoutloudtoanotherperson.VirtualtherapistsystemslikeELLIEarelikeadiarythatcanalsotalkback andrespondtowhatpeoplesay.[8]
4.5 Personalised Treatment
Onethingthatisreallyfrustratinginthefieldofpsychiatryisthattwopeoplewhohavethesameproblemcanreactvery differentlytothesametreatment.Forexample,onepersonmightdowellonacertainmedication,butanotherpersondoes not do well on it at all. One person might get a lot of help from something called CBT. Another person needs to try something else. The field of psychiatry is getting some help from Artificial Intelligence, which is looking at patterns in groupsofpatientdatatofigureoutwhichtreatmentismostlikelytoworkforwhichtypeofpatient,anditisgettingbetter atdoing this. Thisis whatpeoplecall precisionpsychiatry it's liketheidea ofpersonalizedmedicinebutit is usedfor mental health.CompanieslikeSpringHealthintheUSarealreadyusingthisapproachtohelppeoplethrough workplace wellnessprograms.[5]
5. ADVANTAGES
Thebiggestadvantageisthatithelpspeoplereachsupport.MentalhealthsupportinIndiaandindevelopingcountriesis mainly found in cities and only for those who can pay for it. A mental health app is free to download and works on a smartphone.Thatisnotaperfectsolution,butitisbetterthannothing.Andforpeoplelivinginruralareasorsmalltowns withnopsychiatristnearbyitcanbetheonlychoicetheyhave.
Thestigmaissueisabigdeal.Manypeoplewhoneedhelpdonotseekitbecausetheyareafraidofbeingjudgedbytheir doctor,theirfamilyorsocietyingeneral.Whenyoutalktoanappyoudonothavetoworryaboutthat.Youarenottelling anyoneaboutyourproblems.Nothinggetswrittendowninafile.Thismakesiteasierforpeopleto open.Itreallymatters thatpeoplefeelsafewhentheytalkabouttheirfeelings.Wysa'sowndatashowsthatpeopletendtobemoreopenwiththe AIinthebeginningthantheywouldbewithanewhumantherapist.Afterusingtheappforawhileandthinkingthrough someoftheirthoughts,peopleareoftenmorereadytotalktoaprofessional.[3][9]
Consistencyissomethingthat'salsoimportant.Atherapist,nomatterhowgoodtheyare,canhaveanoffday.Theymight be tired or in a bad mood and that can affect how the session goes. They might even have some biases that they're not aware of. A computer system will always give a consistent kind of response every single time. When you are doing somethinglikea guidedCBTexerciseitishelpful tohavethatconsistency.Thebestpartisanappcancheck inwithyou everyday,whichisalotbetterthanjusttalkingtoatherapistforonehouronceaweek.Thiswaytheappgetsabetteridea ofhowyouaredoingbecauseitistalkingtoyoueveryday,notjustonceaweek.
6. CHALLENGES AND LIMITATIONS
6.1 Privacy
Whenyouusea healthapp andyoutell itthatyouhave been thinkingaboutself-harm,oryoutalk aboutsomethingbad thathappenedtoyou,thatbecomesinformationthatissaved.Thisinformationgetsstoredonacomputerserver. So,who canlookatit?Canitbesoldtosomeone?Couldaninsurancecompanyusethisinformationoneday?Mostmentalhealth apps have rules about privacy. These rules are written in a way that is hard to understand, and it is not clear if they are really being followed. India has laws to protect people's information. These laws are new and have not been used much yet.Untilwehavefairrulesaboutmentalhealthinformation,peoplewhousetheseappsaretakingarisk.
6.2
No Real Emotional Understanding
ThetruthaboutArtificialIntelligenceinhealthistoughtohear:theappdoesnotreallygetwhatyouaregoingthrough.It looksforpatternsinwhatyouwriteandcomesupwitharesponsethatissupposedtosoundlikeitcares.Thereisnoreal
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 13 Issue: 03 | Mar 2026 www.irjet.net p-ISSN: 2395-0072
feeling behind it. When you are dealing with something like grief or a bad traumatic experience or a moment when everythingisfallingapart,nothavinganotherpersonwhoreallyunderstandscanfeelemptyveryfast.Abigpartofwhat makestherapyworkistheconnectionbetweenthetherapistandthepersongettinghelp.ArtificialIntelligencejustcannot dothat.
6.3
Risk of Wrong or Harmful Responses
AIsystemscanmakemistakes.Inmostcasesthesemistakesdonotcauseseriousproblems.Forexample,awrongmovie suggestionora badautocorrectcan beannoying but theyare notserious.However,inhealtha mistake canmakethings worse. If a chatbot misunderstands a crisis as something not important it can cause real harm. It can also accidentally makesomeonethinknegativethoughtsmoreoften.Mentalhealthappshaveasystemtodetectcrises.Thesesystemslook forcertainwordsandthenconnectthepersontoemergencyservices.Thesesystemsarenotperfectandsometimesthey fail.Therulesfortreatingthesetoolslikemedicaldevicesarestillbeingcreated.Thepeoplecreatingthesetoolsneedtobe careful.Theymustmakesurethattheirtoolsdonotmakethingsworseforpeoplewhoarealreadystruggling.
6.4 Bias and Unequal Access
Most artificial intelligence mental health tools were trained on English language data from Western countries. This is a problem when someone from a different cultural background uses these tools. They do not work well for people who communicate in a different language. India has a lot of cultural and linguistic diversity. This makes the problem of AI mental health tools even more important in India. There is also a difference in how people access the internet. Many people in India are getting smartphones fast. A lot of people still do not have a good internet connection, or they cannot affordasmartphonethatcanrunmodernapps.[6]
7. SURVEY RESULTS AND ANALYSIS
7.1 Awareness 70%
Most of the 50 students surveyed seventy percent to be exact said they know about AI tools that help with mental healthsupport.Thisseemslikea highnumberatfirst,butweneedtothink aboutwhothesestudents are.Theyarefinal yearengineeringandcomputerstudentswhousetheinternetalotandprobablyknowmoreaboutnewtechnologythan other people their age. If we asked the same question to students in a rural area or to students who do not study technology,thenumberwouldprobablybemuchlower. So,whileitisgoodtoseethatseventypercentofthesestudents knowaboutthesetools,weshouldnotthinkthatthisistrueforeveryone.
7.2 Actual Usage — 50%
Halfofthe people who answeredthesurvey saidthey hadused somekindof healthapp. The difference between people whoknowabout healthappsandpeople whousethem isinteresting. Seventypercentofpeopleknowabout theseapps. Onlyfiftypercentuse them. Thisshowsthatjustbecausepeopleknowabouthealth appsdoesnotmeanthey trustthem enoughtousethem.Somestudentswrotecommentssayingtheydidnotthinkthementalhealthappswouldreallywork. Otherstudentssaidtheirproblemswerenotbadenoughtouseahealthapp.Thisissomethingtothinkabout.Peoplethink theyneedtobereallystrugglingbeforetheycanuseamentalhealthapp.Mentalhealthappsaretryingtochangethisway ofthinkingsopeoplecangethelpbeforethingsgetbad
7.3 Willingness to Use AI 60%
SixtypercentofpeoplesaidtheywouldbewillingtouseanArtificialIntelligencechatbotasafirststepwhendealingwith emotional stress. Most of them said yes but they also had some doubts. This is important to think about. It shows that young people are open to using Artificial Intelligence chatbots, but they are not totally sure about it. They want to learn more before they fully trust these tools. Trusting technology without thinking about it is rarely a good idea and when it comes to mental health it could even be bad for you. Artificial Intelligence chatbots can be helpful but people should be carefulwhenusingthemforemotionalstress.
7.4 Privacy Concerns 40%
Fortypercentofpeoplesaidtheyareworriedabouttheirprivacy.Butthisnumberisprobablynotentirelyaccurate.Itis likelythatpeoplearemoreconcernedthantheyleton.Researchhasshownthatpeopletendtosaytheyarelessworried
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 13 Issue: 03 | Mar 2026 www.irjet.net p-ISSN: 2395-0072
aboutthingsinsurveysthantheyreallyare.Alotofstudentssaidtheywerenotworriedabouttheirprivacybutthenthey also said they would not want their information to be shared with other people. This shows that there is a difference between what people say and what they really think. For people who are making these apps this is something to think aboutcarefully.Ifyoudonotmakeitclearhowyouarehandlingpeople'sinformation,peoplewilleithernotuseyour app, ortheywilluseitbutthey willnotreallytrustit.Thisisnotgoodespecially sincetheseappsaresupposedto behelping peoplewiththeirwellbeing.
7.5 Human vs. AI Preference 30% chose human
Thirty percent of people said they would rather talk to a human therapist than an AI chatbot when they could choose betweenthetwoforfree.TheotherseventypercenteitherlikedtheAIchatbotordidnotreallycarewhichonetheyused. This is because people are used to technology, they know it is hard to see a human therapist, and they might feel uncomfortabletalkingtosomeoneelseabouttheirpersonalproblems.ThisdoesnotmeanthatAIchatbotsarebetterthan therapists. It just means that a lot of people are willing to try a digital tool first when they are feeling stressed. This is usefulinformationforpeoplewhoaremakinghealthservicesforyoungpeople.
8. FUTURE SCOPE
ThenextfewyearswillprobablybringnewkindsofAIsystems.Thesesystemswillbeabletounderstandnotjustwhatwe type but how we sound, what our faces look like and even data from our fitness trackers. For example, imagine an AI system that can hear that you sound a bit down, see that your smartwatch says you have not slept well, and notice that yourmessagesareshortandnotasfrequentasusual.Thissystemwillgetamuchbetterpictureofhowyou'refeelingthan justonepieceofinformation.
Anotherareaofprogresswillbelanguage.NowmostAItoolsworkwellinEnglishbutnotinotherlanguageslikeMarathi, Hindi,TamilorBengali.WhatifAItoolscouldunderstandandworkwellintheselanguages notjustbecausetheywere translated but because they were trained on data from India. This would make a real difference for people in India who wanttousethesetools.Someworkisalreadybeingdoneonthisbutit'sstillnotasgoodastheEnglishlanguagetools.
Ontheprofessionalside,AIcouldbecomeausefultrainingtoolfornewtherapistsandcounsellors.Analysingtranscripts of therapy sessions and giving structured feedback on technique is something AI can do reasonably well and in a country where mental health training programmes are stretched thin, that kind of automated supervision support could genuinelyhelpbuildcapacityfaster.
We really need some rules in place. Right now, anyone can make an app, say it is for mental health and put it on an app store.Thatisnotokay.Weneedtohavesomerulesaboutwhatthesementalhealthappscanpromise,whatkindofsafety featurestheymustincludeandhowtheymustkeepuserinformationsafe.Mentalhealthappsneedtofollowtheserulesso wecantrustthemandusethemwithoutworrying.
9. CONCLUSIONS
Mentalhealthisaseriousproblem.Thenumberofpeoplewhoneedhelpandthenumberofpeoplewhocangivehelpare very far apart. It is hard to see how they can meet in the middle without some kind of technology helping. Artificial intelligence is not a perfect solution to this problem. It has some flaws; it can be risky and sometimes it can even make things worse if it is not designed in a very careful way. At the same time mental health and artificial intelligence is somethingthatwecannotjustignore.Mentalhealthneedsalotofattentionandartificialintelligencecanbeapartofthe solution.
Theresearchandsurveydatabothshowthesamething:youngpeopleknowaboutAImentalhealthtoolsandareopento them. They do not just blindly accept them. They worry about privacy and about whether an app can really help with emotionalwellbeing.Theseworriesarevalidandshouldbetakenseriouslybyanyonecreatingthesetools.
Thebestfutureforusisnotonewherecomputersreplacetherapists.Itisonewherecomputershandletasks,they'regood at beingavailableall the time, beingconsistent,handlinga lotofinformation,andhelpingspot warning signs.Human therapistscanthenfocusonthingsthey'regoodat buildingrelationshipswithpatients,makingdecisionsbasedontheir experience,andprovidingahumanpresencethatcomputerscannotmatch.Toachievethis future,weneedtomakerules for AI tools, train them with more varied and local information, protect people's privacy and keep checking what works andwhatdoesnot.Thispaperisasmallpartofthatdiscussion.
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 13 Issue: 03 | Mar 2026 www.irjet.net p-ISSN: 2395-0072
ACKNOWLEDGEMENT
The authors would like to thank Prof. Tejashri Patil (HOD, Dept. of Computer Engineering, P.K. Technical Campus) for consistentguidanceandsupportthroughoutthisresearch.Thanks,arealsoduetoall50studentswhotooktimetofillout thesurvey yourresponsesmadetheprimaryresearchsectionofthispaperpossible.
REFERENCES
[1] WorldHealthOrganization, "Worldmental healthreport:Transformingmental healthforall,"WHOPress,Geneva, 2022.[Online].Available:https://www.who.int
[2] K.K.Fitzpatrick,A.DarcyandM.Vierhile,"Deliveringcognitivebehaviortherapytoyoungadultswithsymptomsof depression and anxiety using a fully automated conversational agent (Woebot): A randomized controlled trial," JMIRMentalHealth,vol.4,no.2,e19,2017.
[3] B.Inkster,S.SardaandV.Subramanian,"Anempathy-driven,conversationalartificialintelligenceagent(Wysa)for digitalmentalwell-being:Real-worlddataevaluationmixed-methodsstudy,"JMIRmHealthanduHealth,vol.6,no. 11,e12106,2018.
[4] G.Coppersmith,M.Dredze,C. HarmanandK.Hollingshead,"FromADHDtoSAD:Analysingthelanguageofmental healthonTwitterthroughself-reporteddiagnoses,"Proc.2ndWorkshoponComputationalLinguisticsandClinical Psychology,pp.1-10,2018.
[5] J. Linardon, P. Cuijpers, P. Carlbring, M. Messer and M. Fuller-Tyszkiewicz, "The efficacy of app-supported smartphone interventions for mental health problems: A meta-analysis of randomised controlled trials," World Psychiatry,vol.18,no.3,pp.325-336,2020.
[6] S. Chancellor and M. De Choudhury, "Methods in predictive techniques for mental health status on social media: A criticalreview,"npjDigitalMedicine,vol.3,no.1,p.43,2020.
[7] National Institute ofMental Healthand Neurosciences(NIMHANS),"National mental healthsurveyofIndia 2021," NIMHANS,Bengaluru,2021.
[8] H.Gaffney,W.MansellandS.Tai,"Conversationalagentsinthetreatmentofdepression:Asystematicreview,"JMIR mHealthanduHealth,vol.7,no.7,e12185,2019.
[9] Wysa, "Wysa for workplace: Evidence and outcomes report 2024," [Online]. Available: https://www.wysa.com/evidence
[10] WoebotHealth,"Productandclinicaloverview,"[Online].Available:https://woebothealth.com,2023.