This November 2025 P2P Conference plenary explores change drivers, focusing on how emerging technologies are influencing, and will continue to influence, immigration and settlement in Canada. The session looked at advances in AI and digital tools and their impact on: immigration decision-making and processes, the way in which information is accessed and used, and the delivery of services to newcomers.
The speakers were asked to include a future-oriented focus in their presentations, reflecting on how these dynamics may evolve over the next 10 to 15 years in response to the increasing use of digital technology. The goal is to be able to anticipate a number of plausible futures so that we can prepare for them in a more proactive way.
Presenters:
AI-generated transcript:
okay good morning everyone let's get going good morning uh I hope you all had a pleasant evening last night and were able to go to the uh poster presentation down at Pier 21 i know for myself I was able to see a new exhibit that was at Pier 21 which was quite moving and uh very appropriate to have this conference host the poster session there and for many of you maybe enjoy uh a walk along the boardwalk i know that some people were on the search for some famous donuts that Halifax has as well i hope you found them uh but we are going to start this morning with our second plenary of the conference entitled drivers of change in a digital era AI information flows and immigration so I'm going to turn it now over to Louisa Taylor who is the executive director of refugee 613 and the speakers in the first plenary to get us started over to you All right welcome everyone my name is Louisa Taylor and I'm executive director of Refugee 613 we're a communications and mobilization hub based in Ottawa on the traditional unseated territory of the Algangquin Anishinabic peoples and I want to also acknowledge that we are on the traditional lands of the Migmagi people and we're very grateful to be able to gather here today uh on this territory and have the conversations that we're having when Vicki and PTP invited me to moderate a plenary on artificial intelligence my initial reaction was excitement followed by panic excitement because I recognize the importance of this topic and I'm one of those people who loves the convenience of AI you know it helps me draft policies faster find the right document in a sea of PDFs and organize my schedule but I'm also one of those people who feels overwhelmed by the rapid advancement of AI the astonishing pace of its intrusion into every aspect of our work and lives the significant threats it poses uh to equity to ethics and the damaging impact we already see it having in our environment i'm probably not the only person who every time I make the choice to use AI I think about the water being used just for that one prompt I'm giving it so I felt this seductive yet concerning duality of AI in writing these opening remarks and I actually went to AI for feedback on an early draft and it was very helpful but I did not appreciate how it changed my views from what I had put in which was slightly terrified and overwhelmed to occasionally a little nervous so if you too experienced that mix of admiration apprehension fear and curiosity you're in good company me this morning and that's exactly why this conversation matters ai is already shaping immigration policy service delivery and decision-making sometimes in ways we can clearly see and sometimes in ways that are in the background so today's speakers will help us bring those changes into the light i'm delighted to introduce a panel of people who understand this terrain far better than I do and can help us unpack both the promise and the pitfalls of AI in the immigration ecosystem and the impact it may have in the coming decade so this morning's plenary drivers of change in a digital era AI information flows and immigration will look at advances in AI and digital tools and their impact on immigration decision-making and processes the way in which information is accessed and used and the delivery of services to newcomers first we're going to hear from Mario Bellissimo of Bissimo Law Group on concerning implications for the immigration pathways people take to Canada and how their cases are processed then we will be hearing from Elizabeth Coulson Ryan Baloney and Paul Alexander about a joint uh research project that they have undertaken on the use of AI in service delivery and Philip Mai from the social media lab of uh Toronto Metropolitan University will speak to us about information and misinformation on immigration in Canada you can learn more about each of the presenters in their bio in the program we're going to hear from each of them about 15 minutes each and then have time for a Q&A using Slido before we wrap up at 10:45 so with that I'd like to invite Mario to the microphone to get us started thank you everyone it is an absolute privilege to be here and even though this is a presentation on AI to show you how old school you can be here are my notes so not very techy uh but yes but let me begin with a simple question if the gatekeeper uh who's deciding your future uh but you cannot see that gatekeeper how do you trust the process and that's the reality many newcomers face today every every immigration uh system has a gate for decades that gate was opened by a person someone who you could look in the eye ask a question explain your story to many keep many cases that gatekeeper is no longer visible it's an algorithm a triage model a digital system quietly influencing who gets processed faster who is flagged for review who might define you for years and decades to come and those who wait and wait without ever knowing why i see this intersection every day in my advocacy from the courtrooms to the boardrooms to the committee rooms and really what we're asking ourselves is how do we protect people in a world where automation accelerates faster than oversight which brings us to a core tension and we've heard about it already in the conference in moments of high demand excuse me and reducing staff there's a growing allure a belief that technology can make up for the people power and while humans are still in the loop the pathways leading to those decisions are increasingly shaped by digital tools now I believe the technology can be truly transformational not as a replacement for human judgment but as a powerful tool to modernize to equalize enhance and humanize the system when legislated and used responsibly we are not there yet william Gibson some of you may have heard of him the Canadian novelist often called the father of cyber punk i love that and one of the first to imagine how digital worlds reshape real ones once said "The future's already here it's just not evenly distributed." And that's true because technology is shaping not just how applications move through the system but it's shaping how public opinion forms it's shaping how narratives about migrants spread online and how governments respond to those shifting sentiments so today I explore three drivers of change in the digital era ircc's uses of automation digital distrust and automated assumptions and how to best leverage moving forward so ultimately migration is not just about data flow it's a doorway and I like to kind of think about it as every newcomer standing in front of that doorway with something in their hands a degree a dream a child their story so in this digital era we really have to decide what kind of doorway we are building one that scans people in silence or one that opens with clarity fairness and dignity the technology is here as Louisa mentioned it's everywhere in our lives the real question is whether we use it to open doors or to make them harder to see so let's begin IRCC's uses of automation let's look a little behind the curtain so automation use of business rules that are derived directly from the act and regulations and importantly or created by officers based on their experience and expertise those rules we have no access to nor do apparently deciding officers and then some tools use all the things we know about advanced an analytics um and some are combination of all of these and they use these tools throughout the processing it began first with triaging okay and then it escalated keep calling on me oh sorry sorry about that okay so we look at triage and distribution across pretty much most categories now positive eligibility automation some of these terms you may have heard in the past watchtower lighthouse shinook detecting incomplete files responding to inquiries chatbot summarizing case info supporting biometric assessments and these tools can be broken up into three categories decision tools you see there triage tools and processing aids and then there's the more serious tools I would say predictive analytics this ITAT is based on 900 fraud patterns similar to the CBSA's tool which is called TCI a traveler uh compliance indicator it's really predicting your future behavior but we don't have access to behind the curtain then there's some exciting developments the digital moderniz platform modern modernization which you may have heard about this was launched in 2024 and really they're looking to collapse everything into a single digital window for all programs there's an expanded online account that now supports visitor visas and adult passport renewals again a single digital window and then really chat bots client inqu man management systems for faster response time designed to reduce delays and streamline client communication potentially positive developments the digital visa pilot 2025 2026 will issue visas electronically eliminating mailing delays and their testing expansion and then there's AI in the back office operation plan rollout has begun it's continuing and again further expansions with the caveat that the ultimate decision remains with humans thank you so why the digital distrust and what are some of the automated assumptions well from a privacy standpoint IRCC built these all out by privacy by design principles from applicants perspectives there was little to no notice consultation or disclosure on automated uses or AI supported uses and again it became like a treasure hunt what's the newest acronym or newest name heriah Shinook Row ITAT and others like Quaid Lighthouse Watchtower Quantum you may have heard some of these unbelievable and you get all these blogs and people saying "This is exciting what's happening so only after ATIP requests legal challenges and public pressures did we begin to have details released transparency was not prioritized so now the issue is framed as operational secrecy versus legal accountability it's an unfortunate place for us to be now IRCC maintains these systems are internal tools that help officers they do not make final decisions so they do not require legal review to date the federal court has supported this view so now we ask how do we define the human in the loop well now these are this is a decision from a real case reasons began to become quite generic and what you'll notice here is that at the top July 17th 2024 at 1707 is when the officer actually entered their reasons at the exact same time they use Shinook 3+ to populate those reasons so what is the human in the loop is it seconds here it seems to be at the exact same moment the latest iteration of reasons is this some of you may have already seen it um and they're called officer decision notes ODNS which is something we've been waiting a long time for let us see when the when the case is decided what were the reasons for that decision well we finally got them but then all reference through time which officer made the decision if technology was used has all been scraped out we have the generic reasons and you'll see there and any accompanying family member if applicable so let's look at a few real world examples is judging an applicant's intention to visit Canada based on a sibling who made a refugee claim that has no connection to the present applicant a fair indicator well it was used in a trio of cases called KISS an applicant was refused an ETA because behind the scenes quietly a negative indicator was did a family member claim refugee status that's not based in the act or the regulations but it was used and the officer and by extension as I write there might have used their expertise bias that many people who have a sibling that claims refugee status there's a higher preponderance that the individual will claim refugee status that might have been an officer rule not disclosed but used and then what about someone who applies that has been married three times but one applicant is a widowerower and a survivor of abuse onto their third marriage whereas the second applicant has been involved in two issues surrounding marriage of convenience the model may triage those individuals the exact same way quantitatively not qualitatively and then they get sent into what I call orbit so many of our stakeholders in the room have and I won't read these uh because I'm tight on time but they've discussed the human consequences of an opaque system digital divide disproportionate impact techdriven exclusion impacting charter protections which is a live issue now with bill C12 which is a discussion for another day delayed labor market integration economic costs narrow digital windows a programs that open for seven minutes and we refer to those as digital stampedes where it's a click race to get your applicant in no dignity to that we've called for a dedicated AI oversight unit i did it again last week before a parliamentary committee and the guard rails including algorithmic transparency so now how to best leverage the technology and I want to move quickly here because I'm tight on time we need to modernize forms plain language applicant centric technology eliminate the digital stampede guarantee the right to meaningful human review publish full privacy and algorithmic impact assessments not just summaries use model cards to explain and demonstrate how these tools actually function their limits and their oversight develop model card style documentation what does that mean let's see what the officer sees let's see what populates the officer's screen so we understand the journey that our applicants are under anonymize research access communicate officer review times and training how much time are they spending in the Chinook litigation they were very proud to say 120 seconds i don't think that's a badge of honor when people are filing hundreds of pages of documentation and scale individualization and position at IRCC so let me close with just a few comments so every digital tool we adopt we we decide what we will illuminate and we decide what we will obscure we can use autom automat uh automation to light uh to shine a light on opportunity or we can use it in the shadow where assumptions harden and where human stories disappear behind the code ultimately what the pathways to prosperities reminds us all all is that digital systems may be built by government but prosperity is built by all of us advocates universities settlement agencies and every partner who refuses to let these pathways lose their humanity we have a very simple and profound responsibility to ensure the systems we build participate in accept reflect the views we claim openness dignity fairness and shared prosperity we now may be the only human in that loop that newcomers see it's a solemn responsibility and one we must take very seriously moving forward thank you is the presentation on good morning everyone how are you today so my name is Ryan i'm from the Syrian Canadian Foundation we are a charity in Ontario we work with all newcomers and refugees to Canada through three main pillars which are language skill building and well-being today I'm here to talk about a research program that we did as an organization in partnership with the University of Toronto in Missaga and the research is so these are the team members that worked with us on uh this research study um so before I speak what the research is about I just want to discuss if we think about the barriers that newcomers face in Canada what would be a first barrier that comes to mind other than probably employment it's going to be language of course and what's one of the main barriers for language it's mainly confidence and when we talk about confidence we're talking about the confidence to um to speak the conf the confidence to ask questions the confidence to have an accent while people while you're expressing yourself the confidence to ask questions or ask even people to repeat a sentence even if you don't hear it maybe you just nod because you're too shy of asking them to repeat because the other person is speaking fast so because of confidence um this research is exploring how can we use technology specifically virtual reality and artificial intelligence in supporting newcomers learn English faster probably and eliminating the real life anxiousness of confidence so if you're speaking to AI you would ask all the silly questions that you're sometimes very shy to ask right so when we designed this program we are teaching English we developed an English curriculum that actually we developed three English curricula an English curricula for the standard classroom the way we all probably learned at school the second curricula was for artificial intelligence and the third curricula was to be used in VR so when we were designing these this curricula we were thinking how can we design a curriculum that does not only serve research purposes which what we're trying to study but it can also serve newcomers a curricula that participants can connect with they can relate to and they can can also benefits from so it's designed for newcomers specifically who are English CLB level 3 to five because it was hard for research purposes to start with AI with someone who is a CLB level one maybe and we did an emphasis on how can this curricula be practical benefit their day-to-day life and integration and we based it on empathy based interviews and questionnaires so these empathy based interviews that we did prior to developing this curriculum um we got a lot of results from those but there was a lot of focus on the need for community like I want to learn English so that I can express myself I can make friends even someone mentioned that if in this country if you want to drink water you actually need to speak English so the curriculum in mind that we designed was focused on kind of three principles the authenticity empowerment and cultural relevance and I want to highlight here the importance of cultural relevance in that cultural relevance means that whatever client or even language or non- language service we are providing to newcomers it's something they can relate to the more they can relate to any type of services we provide the more the service is successful and the more the positive outcomes that this service has so with these three parallel curricula in VR AI and standard classrooms we had around 140 participants in this study um the students used to come on a weekly basis and attend the classes so this curricula that I'm that I'm discussing we developed mainly these are the topics that we covered throughout the weeks we started with food because who doesn't like food and people connect over food usually during the first session whenever we bring up the topic of food culture what's your plate people connect and I'm again highlighting the word connect because with technology or without technologies this is something that we saw throughout the program the need to connect with other people was always there so we started with food and then the the topics we covered related to Canada cultural famili familiarization so um we all got culturally familiar in Canada in our encounters some encounters were nice some encounters were challenging or uncomfortable so what we tried to do is actually have this cultural familiarization in a safe um in a safe place kind of so for examp what is an example of cultural familiarization um I had one newcomer he was a week in Canada maybe and he was working at Tim Hortons and some person came a client just ordered coffee and the p the client asked to give him back looney and tunis so the Tim Horton person didn't understand that like loonies and tunies so the person repeated "Yes I want loonies and tunies." So the person went to his manager and told him "There's this man here asking me about a TV show Looney Tunes." So that's exactly cultural um familiarization right and I feel it's our job as settlement agencies as as nonprofits who work directly with newcomers to kind of make this process less intimidating less challenging less uncomfortable for the newcomers and it could be done in very little small tweaks here and there so banking and finances we did the English class was about all the financial terms that one can use in the bank like debit card credit card checking account and all these terms that they can learn and this these terms will actually help them in their day-to-day life as they go and open a bank account they've explored that they've experienced that um health and well-being employment uh mock interviews so what we did in the VR for example all the students in the classroom were in the VR in a VR space that we created so the session on banking and finances it was an office where there's a bank teller and students were having this conversation virtually and this actually removed the in-person element which is why one person mentioned even when on VR I feel nobody is seeing me and it helps me speak so the no one is seeing me is actually a push for people to make mistakes and it's fine i'm not I'm not there and there's not the real life kind of stress or anxiety um so the future of language learning for newcomers when it comes to technology um I also have my notes on paper um technology is very helpful and it does eliminate a lot of factors that newcomers face like the stress of speaking um the anxiety of making connections but what we did notice in this study is the importance of community for people to actually feel comfortable and to thrive in general um AI is a tool and VR are tools that kind of eliminate those factors but there's no replacement of the human design element so what made the study successful is having these components of the curriculum which tackle everyday life and where newcomers can connect with this is what made VR successful it's not VR eliminated the confidence and things just worked and I feel that this is where um the human design customization aspect is required and then AI would or VR would be the tool that actually help miticate other factors that just us humans probably it's more helpful than on person so this is it for me thank you good morning everyone uh my name is Dr paul Alexander i teach educational technology and immersive technologies at the University of Toronto and I'm just going to add on to what Rajon was sharing with you and if you look at the slide what I have here is just uh a a quick graphic or two to uh illustrate the the design of our research studies so what we had was we would have students come into our our campus uh once a week over a two-month period and we would uh randomly assign them to one of two classes there would be a standard class where they would just learn English as they normally would anywhere uh say at the YMCA uh and we also had another class that was VR centric and so what we would do is the students would come in usually on a Saturday and we would uh the the the whole the whole day they they would be there for about half a day and so their lessons were almost two and a half hours long so what we what we would do is the first part of the lesson we would just teach students uh as we normally would in in a regular class we teach them vocabulary grammar sentence starters uh and get them uh talking about certain topics that Ryan uh spoke to banking uh small talk maybe job interviews things of that sort um but you know if anyone here has ever taught uh English language before you know a big part of the the lesson usually involves the roleplay aspect and so this is where after they've learned all the vocabulary and grammar now they put it to practice and in a normal classroom usually we just uh meet up with someone that a classmate sitting next to us our elbow partner and we would just take turns uh going through a script and and that's fine and so one of our classes did that but the other class they were involved with the the VR headsets so uh when it came time for them to do the role play rather than just talk with someone in their their classroom we would uh assign them a headset and they would go into this virtual world where we had set up these scenarios for them to to practice their language um you'll notice one thing is that the the instructional part was uh usually about 60 minutes long and then it was followed by 20 minutes in VR we were very cognizant to make sure that our students because uh our Canadian newcomers none of them or almost none of them had ever used VR headsets before we didn't want to overwhelm them uh with with this technology so we we made sure that their role playing uh uh moments were only about 20 minutes long and we felt that worked quite well and then we'd take a nice lunch break or or break and then come back and repeat the same strategy um just very quickly some of the takeaways uh we felt that uh pedagogy or teaching methods was really really crucial to having success in this in this study um teaching in in just a traditional class for example uh when students go into VR they tend not to listen to teachers uh when they go into VR they're in a space they're like little kids they run away and they they like to explore so what's really interesting for teachers by the way when you're in these platforms is there's this one thing in a teacher's menu system and it's there's this big red button and the big red button has a one word on it it's called summon and so whenever the teacher wanted to bring the students back to one local place in VR just click that red button and it would just randomly collect uh collect all the students in that VR space and just bring them back to the teacher so we thought that worked quite well um this is just a quick rendering of our classroom you can see uh you know it was nice and spacious and we would we would set up the students at different tables when they were in the VR experience we quickly discovered that there need to be more distance between the students so we didn't have the audio feed feedback uh happening um moving on here here are some screenshots from some of our scenes uh the first lesson or two was really about learning how to use the technology how to use the hang controllers move about teleport walk and then even in the right image turning their virtual wrist over where a little menu would pop up and they could grab a 3D pen and start drawing or or type post-it notes and so forth um here are some screenshots from some of our scenes throughout the program uh we had a meet and greet in a gallery in the top left corner uh we took them to an island where we worked on some of our adject uh adverbs of frequency on smart on whiteboards um in the bottom left corner um Radian talked about uh Canadian terminology like loonies and tunies and maybe double double and things like that so we had a a lesson in a cafe and maybe the most impressive uh VR experience was the job interviews the students had actually pra uh prepared cover letters resumeumés and then they'd come to our class and our teachers would uh interview them in a in a little boardroom and they were so excited their heartbeats would just not stop beating um so rapidly uh very quickly uh just another minute um I just want to share a few uh quotes from our students the top one is is rather long but essentially it speaks to a student's reflection where she thought that she had gone through many English programs and she thought that the VR experience was very authentic and meaningful her uh meaningful for her um she she thought that we had tied the language to uh real life experiences that they might have in their everyday lives um the the one in the bottom left corner one thing that was a nice takeaway was when students were in virtual reality and they had these cartoon avatars uh they felt this sense of freedom to speak uh they weren't so conscient anxious about their their pronunciation or accuracy with their verb tenses they just wanted to talk and so that was a nice uh nice takeaway and then the teacher in the bright uh bottom right corner he speaks to the engagement that students had with each other and with the technology and the excitement that came from it um I do have uh further slides to do with uh stats but I'm going to save those for uh the conference if you want to look at them later uh but one of the the big takeaways for us was there was a lot of uh preparation in terms of using the technology um and then also one of our big things was teaching teachers how to use uh the teacher training aspect to to uh guide them in teaching students effectively in these virtual worlds and I'm just going to move on to Professor Liz she'll come up and talk a little bit about a parallel study to do with AI hello everybody good morning wonderful to be in Halifax and uh thank you to my co-presenters as well um I just wanted to share with you the second year of this project and I'll do that very quickly because I know we're short on time um but we actually in the second year we transitioned out of VR and we started using AI so we had an experimental group that used AI i'll just flip to a slide so we could get started here sorry one second back yeah so we did a quasi experimental mixed method study to test the use of AI in one group and to test the uh non-use of AI in our standard group um and what we wanted to do was involve university students as well so what we did was we invited in uh we trained university students to work in our standard class and then we had AI as our tutor in the in the AI class so I'm going to just tell you a little bit about that project um so we had the same themes running through the AI program that Paul explained to you and we used we first of all developed priorities around what we needed to see in our chat bot or the we used a priority a way to check and see what we wanted to use we ended up using chat GPT it was just coming out rolling out at the time and uh we wanted to use language learning so we decided we would use the oral communication tool within chat GPT so students were in a computer lab and basically they were set up to use prompts and we collected the data from the prompts as well which will be future analyzed but I'll talk about what we did analyze um so we used something called uh EIT so an elicit imitation test and what we did was we provided 20 sentences that related to each of the themes and the students in the first week of class actually listen to each of the sentences read aloud and they had to respond on their phones to what they heard from that sentence so it's elicit imitation response assessment tool and we looked at the data at the beginning and the end of the program for both cohorts so those that were in AI the AI classroom and those that were not in the AI classroom in the standard classroom we were trying to investigate if our human tutors in the classroom had a more significant impact than our AI tutor in the classroom and just to kind of summarize very quickly I'll just give you some information here um we used a multi-level modeling tool to analyze our data we had our students helping with that as well but we actually found there was no significant difference in the two develop in the development of language skills in the two different groups so they equally moved forward in their language skills uh the nonAI group control group and the AI group experimental had equal impact in terms of development of their language skills so you can see from this slide here as well um so we were kind of a little bit nervous about what do we do is the teacher being replaced um how do we how do we deal with this information when we when it comes to AI and uh as we move forward with the study whoops sorry I'll just go back here a bit um we started to think about ways that AI can be used um as a co- u co-guide along the way and we started to look at um elements of how we can change our change and improve our content based on that experience so we looked at the study from an AI perspective a VR perspective and you can look a little bit more at our data in our slides but I won't take any more time because I know we're rushed thank you all okay so good morning everybody uh my name is Philip Mai uh I'm going to switch gear on you a little bit and talk a little bit more about the darker side of AI and how it's being used to basically fuel anti-immigrant um misinformation here in Canada and abroad so I'm going to start with a little bit of a quiz for you which of these photos do you think is AI generated i'm going to give you a few seconds to make up your mind okay actually um but as you can see it's very very difficult now in the early days when some of these um technologies were released you could check the fingers you can check the eyes and a few other things but over the last few months they've gotten much much better so that what that tells you is that we're about to be hit with a wave of misinformation in the future where it will be very difficult for us to make sense of what's going on in the world and to figure out which way to go on issues um and that's going to affect everybody here in this room and everybody you know so to start with I just want to tell you a little bit about the lab that I am from so that you know where I'm coming from and um talk a little bit about how Canadian consume news and how Gen AI adoption Canada is the a concern for Canadians and talk about a little bit of the risk and mitigation strategies for immigration and settlement in particular i'm part of uh the social media lab at Toronto Metropolitan University we basically look at how people use different technologies like social media AI and so on to you know live their lives do their work and be part of the Canadian society over the last few years we've been doing a lot of work on propaganda misinformation foreign interference and so on so for example we recently last year looked at um what type of misinformation from Russia and disinformation from Russia do Canadian believe so we've surveyed 1500 Canadians we discovered that you know individual who holds uh right-leaning ideology are more likely to believe in pro- Kremlin disinformation for example and those who believe in the claims are more likely to rely on social media not surprisingly because if that's where they're spending all of their time that's where they're going to run into a lot of this type of content and the other thing that we just did recently was we were looking at how um Ukrainian refugees are being talked about and discussed if you think about it the Ukrainian refugee crisis is something that is unique in recent memories it is the first time where the refugee group is majority white every study on um refugees and immigrants in the last 50 years focused on people of color so what we wanted to know is how does the bad actors online who are um maligning refugees as a whole but how are they attacking this particular group without being able to use some of the tropes the skin colors the language the religion uh wedges that they normally use how do they talk about this particular group to convince policymaker not to support these people so if you're interested in that study the um QR code is there i'm not going to go into over into detail because that's not the purpose of this particular talk the other thing that we do at the lab is that we publish a lot of different public reports that um help shed light on you know where people in Canada are spending time for a long time a lot of time when a reporter calls us and asks what are Canadian doing online we often say look at Pew Internet in the US uh and whatever the Americans are doing we're probably doing that but after a long time of saying that um me and my colleagues look at each other like well if anybody in the Canada is going to be doing this longitudinally it's probably going to be us so we have to figure out how to do these studies every two years so every two years we ping 1500 Canadians and we say "Hey what are you doing online where are you hanging out where are you you know what platform are you abandoning what platform you're picking up so you can find all those different reports online uh right now and the other thing for the researchers in the room um we also make software to help researchers and academics get the data they need because not everybody wants to go out and get a second PhD in computer science just to get the data so what we've done is uh we made a tool so that if you know how to use the search bar you can then get the data set that you need to do your study we can get data from Blue Sky Mastedon Reddit Telegram uh even X if you're willing to pay Elon a little bit of money for the data uh but once you pay him the money you can come to us and use our tool to get the data that you need and also you get from YouTube and in January we're going to release a tool that will let you get data from Tik Tok also and the other thing that we do is we make dashboard to track misinformation so for example when the um Russia Ukraine war happened overnight in 24 hours we created a dashboard to track all fact checks around the world involving the Russia Ukraine war so instead of you going to Reuters or going to AFP or um a news site in India to see if they've done any factchecking relating to the Russian Ukraine war you simply go to the conflict misinfo or we go to each and every one of those news site from around the world we look to see if they've done a fact check recently involving the Russian Ukraine war we translated it into Russian Ukrainian and English and bring it all into one dashboard so in the morning you just fire up our dashboard and you can see all the latest fact checks involving the the war and another app that we made is the defake tracker again same thing if there's any fact checks by any news organization around the world that involve a deep fake we would bring it in so you again you don't have to go to each of these sites to say hey do you have anything lately about defects and note tracker is something for those of you who are still on X um this is a thing that tracks community notes community notes are user submitted fact checks right but the problem with community notes is that they rarely ever see daylight the majority of community notes are proposed but they don't meet the threshold to be visible what we've done is we gather all 1.7 million community notes and we make it available to you so if there's a topic that you're passionate about or you're monitoring and you want to know if there is uh community notes related to it you can come to our app and you can surface that particular community notes so you are aware of the conversation around that topic so in summary what we do is we research we train we organize events and attend um events like this and we develop software to help other people do what they do so next I want to talk a little about news consumption habits and social media use in Canada this is critical because as you are trying to reach your stakeholders not only the people that you serve but the people that are going to be funding you the people that are going to be helping you and joining you at your next event um you want to know how people are consuming news we want to know where what social media people are using in the old days people go to the town square they go to the mall um they go to church they meet and talk and discuss what they like to do together next now that's happening online um and as we always say in the lab we go to where the people are and right now the eyeballs are online so we have to be online so uh Reutder institute just released a latest report where they show that most Canadians uh 72% get their news you know online and u via or via social media 46% get from social media but when you combine um you know social media and um digital stuff it's a whopping 72% literally that is where people are spending their time so that's where you need to be if you want to reach a larger audience and a more targeted audience um I know that some of you might not like that idea but sadly that is the reality that we face um one of the report that we put out we do this every two years we've been do this is the fifth u time that we've done this um where we like I said ask Canadians where they're spending our their time online surprise surprise it's still Facebook you know that's the reason why is people want to know when the next birthday party is they want to know which party they were not invited to um so that's where it's at they may not actively post on there but they're still lurking i know you are um surprisingly for the the latest version the one platform in Canada that's actually losing user is X since Elon bought X he've changed the flavor of what is online on that platform and as a result people are abandoning the the platform maybe not to the level of some of us in the room would prefer but there is a noticeable drop and we were able to document that and next I'm going to talk a little bit about another survey that we uh just released on general uh AI adoption in Canada and concerns about misinformation as you know many of us are having AI thrust upon us if you go to work you open up and sign into your system and one day and your boss announced that hey there's a new AI system that you know we prefer you to use uh you go online you use your favorite um browser and tada there is now an unasked for AI feature uh and if you don't use it it nags you it's like that little clippy uh thing from Microsoft a long time ago do you want me to help you with this that's what it's there so it's hard to avoid but the reason why it's everywhere is that some aspect of AI is useful it is useful for some task but it is not necessarily used for every task um and one of the things that we as a society have to figure out is which of those tasks is it good for and which one we should not let it touch ever as Mario was pointing out earlier there are certain things that we need to make sure and insist upon having the human in the loop one of the questions we ask is like do you have you used an AI tool you know 66% of Canadians that they have right um the entryway for most Canadians is through leisure they're doing it for fun it's fascinating it's interesting let's see what this thing can do let's see how it mess up all right that's what people are looking at um a lot of the people who are using it for work and and study are actually the younger generation that tells you a lot it tells you that as they finish school when they go into the workforce this is something they're going to take with them so whether we like it or not we have to figure out what are the processes and procedure that we need to put in place now so that when these things are being used we're not surprised by their outcome or their unintended consequences and the other thing we ask is to what extent do you agree with some of these statements so for example we we want to know whether they know how to use these tools a lot of people don't right it's not surprising it's so new and literally it's changing all the time i remember in 2023 when we were um asked uh by government department to write up um the uh progress of AI and how it's going to change democratic institution in Canada we were literally changing the chapters week by week because when we wrote something the week before we said it can't do that yet the week after it's like well actually it can't do that now so we have to go back constantly that's how fast things are changing and lastly we want to know like what do they worry about and a lot of people are concerned about future elections simply because when these are these tools have become more ubiquitous I don't know what the politician is targeting and sending to you to get your vote versus what messages he or she is sending to the other person standing next to me in the old days somebody like me standing up on a podium like this saying something each one of you can turn to each other and says "Do you agree with that?" or "Oh I agree with that." Or "I don't agree with that." Right but now when we're own little private little world with our own little phone I have no idea what that politician is telling you to get your vote i don't know what telling your neighbors to get his or her vote and that's the problem that's why targeted advertising is quite dangerous and lastly there's a recent report from StatCan that shows that a lot of Canadians are indeed worried about um misinformation so this is something that's not going to go away this is something that's part of our the bloodstream of our EOS information ecosystem today so last I'm going to go into risk and mitigation i might go over by a few minutes but um the recent um landscape in Canada some of you if you're online you probably seen some of these uh AI images if you notice some of these look quite fake and you and I could see that but the point of these images is not to make you believe them but it just helps to change the conversation it forces you to think about these things it forces you to talk about them and it's used as a recruitment to identify other people who believe this way so that these groups can find new members right because people who will join in the comment section who will say things that says they support this guess what they will be privately reach out to join so as we sit and watch thinking that oh anybody I mean nobody could believe this this look obviously fake that's not the point of these fake AI it's not to make you believe is to find like-minded individuals so that they can grow their movement okay so that's another reason why they're using them but one of the reason why it's so dangerous it's so cheap now like in the old days you have to have somebody with specialized skills to make something that looks good that looks believable and it'll take time now literally if I can type something into a search box I can get an image that is passable it's a lot faster right um is more visual and more targeted i can create a whole bunch of different images targeting different groups of people so now in a day I can create 10 15 ads right showing uh you know an Asian face showing a a Muslim um person in the same ad changing very little but now I have created all this content that I can use to target different people to make them believe what it is that I want them to believe um there's some examples that you might have seen um here in Germany one of the far-right groups they specializes in this they literally flood the airwaves and online with this stuff um this summer when there's that tragedy with the Southport stabbing that causes riots all over the UK within minutes of the incident somebody created that image a very highly charged image uh with Big Ben there um and a child screaming uh and bearded men running through and look how much um attention it got within minutes right so this stuff can go viral in second in the old days it would take them time and by that time you know things have moved and changed but now they literally can join the the conversation in minutes so negative impacts on jai on immigrants and settlement this is the meat of my presentation for immigrants it's a safety and belonging situation where they don't feel like they belong if they're constantly being bombarded and being told that they don't belong u because of this type of content service uptake um they may not want to show up at some of your physical location if there are protests in front of it um mental health situation scam and impersonation they could be be target of scams hi you got a message from immigration services hi you got an office from the settlement services um and so on so impact on service is real uh there's lots of operational risk um your organization could be the subject or the target of you you know take an interview from a reporter a day later somebody is now become aware of your existent they could then create misinformation about you and target your group um and that could hurt the ability of you to get funding in the future for example but more than anything else more immediate could be the harm to your staff um people can vandalize your um office and so on so some resilience and um um mitigation uh strategies you need to start thinking and talking about establishing an AI policy in the office when do you use it when you shouldn't you use it how do you mandate that there is a human in the loop so that somebody sign off on it and not just phone it in and you know the AI did it and just shoot it out the door there needs to be a process in place where you say somebody verify this before you move on with it and then the other thing is find people in your staff who are more likely to be the target not everybody in your organization are going to be as equally vulnerable but some staff members will be targeted simply because they are more public-f facing so provide additional training identify those people and provide training for them um and proactively warn your client when you hear of things that might affect them don't wait for them to come in and says "I see this is this real?" Keep an eye on those things so that you can proactively send out messages that says "Hey this is what's going around in our community." Start WhatsApp group with each other share things and incidents so this way you're aware as a community and before I go I'm going to give you a peek at something that we're working on right now that we're going to release in February as you know a lot of people are using Facebook ads to run ads that are potentially you know anti-immigrant and it's a difficult to find them until they make um a lot of noise and cause their um um impact on society already what we want to do is create an app that as soon as somebody published an ad that's anti-immigrant our system will be able to flag it and then if you uh we you know organization can then get together put out press release put a spotlight on these bad actors um so that as they're running these ads we will know them as soon as they run them as opposed to waiting for the ads to um go viral and have an impact by then it's too late to put out that press release okay so thank you all right we are going to move to questions in just a moment and I believe we're going to have a slide um in uh shortly to show you how to log on to Slido but first um I want to say a huge thank you to all the panelists i was furiously taking notes learning so much from each of you thank you so much um so um we're going to Okay there's the slide out while you're sending your questions in I would just like to ask the AV team um we're going to show you a video because this session is sponsored by Pre-arrival Canada YMCA of Greater Toronto so while you're voting on questions we have a short video from Pre-arrival Canada before moving to Canada you must have a lot of questions we are immigrating to Canada and we don't know where to settle down will I find a good job before I run out of savings would anyone help me transition to school hey guys here is what you need to get ready before moving to Canada settlement programs will prepare you for your new life in Canada they'll provide you information on finding accommodation and registering children for school employment programs support you with your job search writing Canadian style resume and learning about the workplace culture and we can help you to find a mentor to guide you you and your family can get free online services pre-arrivalcanada.ca is designed to help you prepare before you arrive don't wait visit pre-arrivalcanada.ca and register today pre-arrivalcanada.ca DA and all programs offered are funded by IRCC Immigration Refugees and Citizenship Canada okay great we did go a little bit over time with the speakers uh but I don't think anybody here would be complaining it was all really really valuable so let's start with a very popular first question how do we balance advocating for more human oversight on immigration application reviews with clients demand for faster processing times and the federal's budget's emphasis on AI and productivity i'm going to give that to you Mario um thank you you know in my view I I'm not certain that they're mutually exclusive uh one of the things we've advocated for is some guard rail legislation to at least frame the core essentials explicability uh redress uh the ability for an individual to have a human alternative if not uh eligible or have access to certain um technologies um faster processing times this was kind of the sneaky quiet under the radar result of the heavy demand we saw in 2022 they processed over 5 million decisions IRCC which was nearly double the year before and it's escalated since then so the need for faster processing cannot come at the expense of individualization or fairness um and there is a way the technology allows us to balance both um it's getting the legislators um to appreciate that context the courts to understand it's not the technology is not the focus really it is the assumptions and the human talent um that underly that technology so that's how we have to look at balancing it we need some guard rails we need some essentials and then build out from there all right thank you and one for for you Philip I think how does the news ban from Canada's online news act impact disinformation and potentially the promotion of unbacked opinionbased reporting on social media i have not done seen any study that's shown the exact impact but I believe that it is having an impact due to the fact that right now when people are on Facebook they cannot share links to any news media so if they see something that they uh think is incorrect there's no way for them to share or post uh anything that's reputable to refute that um so I think that's something that we have to work on and get that changed all right and I'm just going to uh go down a little bit to um one for the researchers on the panel uh the the English language training as AI becomes more common in teaching English to newcomers how do we assure that if it ever replaces teachers immigrants still get the human or emotional support they need during that long journey yeah I think it's a great question that we're all grappling with one of the things that uh we're trying to do is look at teacher training and where we're headed in that direction and uh basically w with our project we started a new u kind of training environment for that looking at tools and activities that are integrated u but I'll pass it to Paul because I know you've had some experience directly in your class with your AI students yes uh so we're trying to figure out ways uh where we can use AI in efficient and ethical ways and uh in some of my classes I've done research studies where we use AI as a uh as or Gen AI as a AI chatbot where we have students uh log into a certain platform that we've developed and we use it for students to demonstrate uh their understanding of of course content um it seems to work quite well um it's the they the the students are always trying to hack the system they're always trying to find g give the answer uh that the AI chatbot is asking uh we've made it pretty much foolproof so that students have to spend about 15 or 20 minutes interacting with this AI chatbot and basically coaching them or teaching them uh what they know about certain subject matter uh with regards to uh replacing teachers I mean that's I think that's a concern in so many industries uh I don't know if we have a real answer at this time for that but uh that's where we're at with AI risk and opportunity everywhere we look really all right well by far the most popular question is what is the federal government getting right about AI adoption and what is it getting wrong what can or should Canadians do to make sure the feds get it right and I'm going to throw that open to the floor don't be shy i can start um I think the directives they've come out uh with uh it's actually one of the leaders in the world so we we've done some good what I call soft law to give us some guidance the Treasury Board directives the Geni directives which are positives um I think the way and IRCC is at this point my understanding the largest user of this technology um so they've they've been a leader um and there the algorithmic impact assessments are encouraging but you can't have you know you don't uh shouldn't need a degree to be able to understand them um and that's part of the plain language and and Philip can probably speak uh to this better than I can but um the issue of deep fake uh especially in our world now as advocates or support is is the evidence that's being brought forward uh by your own the the newcomer that you're dealing with is it legitimate and can you rely on it and we do not have AI detection tools that are ubiquitous and again Philip I'll throw this to you but I think that's one of the things that the federal government really has to grapple with is how far uh what devices the federal courts come out with an AI declaration which says when you're you know as an advocate you're using uh AI in your pleadings to uh declare that uh but it is a real challenge yeah so I'm going to touch on that a little bit the main issue is transparency people want to know that you are using that if you are you should declare it why if you declare it it makes people more aware and then they can question the results of what you are deciding on and the other thing is that we need to be teaching critical thinking and and go beyond just media literacy um remember when Google first came out librarian dropped everything and basically spent well four or five years just teaching us how to Google better right we need similar initiative now to basically teach people the limitation of AI what it can be used for what it can't be used for um I know the technology is new and is still rolling out and new features are being added but this is something that we as a community needs to be staying on top of because the people that we are um helping are going to have to grapple with these and they'll have to do it on their own and we're going to be the kind of people that you turn to so you need to be starting having speaker like us coming to your events to talk to your community for example but also when you are able to talk to uh politician remind them that you know people still don't know from our survey you can see that a lot of people still don't know how to use these tools right and everywhere I look I haven't seen anybody putting out any fundings at all to improve that kind of training thank you one last question before we wrap up and I think it's important on a a serious topic like this to give people maybe a sense of a way forward so how can IRCC funded organizations like many of the organizations in this room but I I would I would argue it's not just IRCC funded any organization working in the sector supporting immigrants engage in advocacy around AI and IRCC or government usage of it where it impacts immigrants are there ongoing campaigns that organizations can sign on to as a way to mitigate the advocacy funding limitations is there a working group that we know of that is focused on AI in immigration so I I can speak to um there are RAI working groups at the Canadian Bar Association and various organizations and I think as settlement agencies uh it's important to collaborate with those organizations uh because ultimately uh it is going to be about delivery it's going to be about effectiveness and it's about insisting on certain as I've said before guard rails that are essential in the work that you do you should be able to have that that explicability that transparency that Philip spoke to and these are the red lines we need to insist on not only when we're dealing with our own clients but with other organizations and I think coming together as I said in my opening remarks collectively it's a profound responsibility right now you know we're talking about generative AI but we're already on to a gentic AI and that is a further dependency uh and a further uh move towards ultimate reliance on people believing this is an oracle as opposed to a support so I I think there's a key role to play thank you with that we are going to wrap up i want to say a huge thank you to all of our speakers you've given us so much to think about and I think I'll I'll leave you with this parting thought that that came actually Mario uh from your presentation is thinking about what kind of doorways we're building and whether they are visible or invisible we all deal with these doorways in different forms um and uh as you move through these sessions today you can think about the drivers of change which is the theme of this whole conference consider the opportunities offered by ARI and the safeguards needed thank you all so much i believe there is a break now and then it's on to the first uh breakout sessions of the day thank you
Subscribe to get the latest posts sent to your email.
Please take this short survey to help improve the KM4S web site. The survey is anonymous. Thank you for your feedback! (click on the screen anywhere (or on the x in the top right corner) to remove this pop-up)