GEF Madrid 2024: Conversation: Becoming an AI University / GEF AI Platform

Summary

George Simons, a co-founder and chief scientist of SNH use human system, discusses the intersection of human and artificial cognition in education. He addresses the challenges and opportunities AI presents to higher education, emphasizing the need for universities to adapt and integrate AI to remain relevant. Simons outlines key areas where AI can transform learning, teaching, and research, urging a proactive approach to ethical considerations and data infrastructure.

Takeaways

• 🧠 The speaker, George Simons, researches the intersection of human and artificial cognition in knowledge processes and is involved in building resources to respond to the impact of AI on learning and wellness.
• 🏛 George criticizes higher education's response to AI, suggesting it has been fatigued and even antagonistic, and calls for universities to be more proactive in utilizing AI to ensure student success and remain relevant.
• 📚 He discusses four main topic areas: literature on AI and learning, the intersection of human and machine cognition, current trends in AI, and the implications for universities.
• 🤖 AI is not just a tool but a node within our cognitive network, which has significant implications for education and how we understand and facilitate learning.
• 🔬 Research indicates that AI can affect complex problem-solving and learning processes, including metacognition, emotion, and trust, necessitating a re-evaluation of educational theories and practices.
• 📈 The benefits of AI in education include personalized learning, improved administrative processes, and more effective assessment methods.
• 🚧 However, there are challenges such as ethical concerns, curriculum development, and infrastructure questions that universities need to address when integrating AI.
• 🔍 A literature review reveals that AI applications in education are primarily focused on adaptive systems and personalization, profiling and prediction, and assessment and evaluation.
• 🛠️ The current trends in AI include the prevalence of AI in various platforms, the growth of open-source LLMs, and the integration of AI with traditional robotics and wearable devices.
• 🏗️ To effectively incorporate AI, universities should focus on building institutional capability, improving personalized learning experiences, and accelerating research through AI utilization.
• 🌐 The speaker advocates for a global data consortium to facilitate multi-institution collaboration and sharing of data to enhance AI capabilities across universities.

Q & A

What is the main topic of George Simons' discussion?

George Simons discusses the intersection of human and artificial cognition in knowledge processes, the role of AI in higher education, and how universities should adapt to remain relevant in a rapidly changing world.

What is the significance of the speaker's role as co-founder, chief scientist, and architect of SNH use human system?

As a co-founder, chief scientist, and architect of SNH use human system, the speaker is involved in building resources to respond to the systems' impact on AI, learning, and wellness, highlighting his expertise in the field.

What are the four distinct topic areas that George Simons plans to cover in his talk?

The four topic areas include literature around AI and learning, the intersection between human and machine learning, current trends in AI, and the implications for universities, as well as priorities for universities to consider in the AI conversation.

How does George Simons view AI in the context of human cognition?

George Simons views AI not as a tool, but as a node within our cognitive network, suggesting that AI is an integral part of the networks that make up human intelligence.

What is the current state of AI in education according to the literature mentioned by George Simons?

The literature suggests that AI in education is primarily used for adaptive systems and personalization, profiling and prediction, assessment and evaluation, and to a lesser extent, tutoring.

What are some of the challenges introduced by AI in the educational landscape?

Challenges include ethical dynamics, ensuring AI helps rather than harms people, preserving student integrity and security, curriculum development, and infrastructure questions.

What does George Simons suggest as a critical component for universities initiating the AI conversation?

George Simons emphasizes the importance of getting data and related architecture right, building institutional capability with AI, and considering multi-institution collaboration for sharing data.

What is the concept of an 'AI first University' as proposed by George Simons?

An 'AI first University' is one where AI is involved in all aspects of the organization, from infrastructure to admissions, teaching, assessment, curriculum, and the research process.

How does George Simons relate the impact of social media to the potential impact of AI on society?

He draws a parallel between the initial positive connections provided by social media and its eventual negative effects, suggesting that AI could have a similar trajectory if not managed properly.

What are some of the significant trends in AI that George Simons identifies?

Significant trends include the prevalence of AI in everyday technologies, the growth of multimedia and multimodal AI, the promotion of open source LLMs, the rise of very small LLMs, AI pairings with traditional robotics, and advancements in platform technologies for AI development.

Outlines

00:00

🤖 The Integration of AI in Higher Education

George Simons, a co-founder and chief scientist of an AI-human system, discusses the underutilization of AI in higher education. He emphasizes the need for universities to adapt and innovate using AI to ensure student success and institutional relevance. Simons outlines four key topic areas: literature on AI and learning, the intersection of human and machine cognition, current AI trends, and the implications for universities. He also addresses the societal and emotional impacts of technology, cautioning against the potential negative effects of AI on mental health and societal well-being.

05:01

🧠 Embodied and Distributed Cognition in Learning

This paragraph delves into the concept of cognition extending beyond the brain, with theories of embodied and distributed cognition suggesting that intelligence is a function of the networks we exist within. Simons argues that AI should be viewed not as a tool, but as a node within our cognitive network, affecting educational practices. He discusses the impact of AI on metacognition, emotional regulation, and learning management, and the necessity to re-evaluate human learning and knowledge growth in the presence of AI. The rapid decision-making capabilities of AI systems and the challenges they pose to human intervention are also highlighted.

10:02

📚 Practical Applications and Challenges of AI in Education

Simons reviews the practical applications of AI in classrooms, focusing on adaptive systems and personalization to provide a one-to-one learning experience. He also touches on profiling and prediction to better understand student capabilities and assess their progress. However, he points out the ethical challenges, curriculum development needs, and infrastructure questions that arise with AI integration. The research gaps in the literature are identified, with ethics being a prominent concern, and the need for new methodologies in educational research is emphasized.

15:03

🌐 The Evolution of AI and Its Educational Implications

The paragraph discusses the evolution of AI and its impact on education, noting the growth of generative AI and the practical applications emerging from it. Simons identifies trends such as the prevalence of AI in various technologies, the rise of open-source LLMs, and the integration of AI with traditional robotics. He also mentions the importance of platform technologies that simplify AI development and the potential for AI to impact all aspects of a university's operations, from knowledge generation to student recruitment.

20:04

🛠️ Transforming Universities with AI: An AI-First Approach

Simons proposes an AI-first approach for universities, where AI is integrated into all organizational aspects, from infrastructure to admissions, teaching, and research. He outlines six areas of focus for universities to consider when engaging with AI, including data architecture, building institutional AI capability, leadership and policy, adaptive teaching methods, and accelerating research through AI. The importance of getting data and architecture right is stressed, along with the need for multi-institution collaboration and sharing data to learn from peers.

25:08

🌟 Reflections on the Future of Education and AI

In the final paragraph, Simons reflects on the broader implications of AI for education, moving beyond teaching knowledge to developing human beings who can navigate complexity and engage with non-human intelligence. He emphasizes the importance of proactively addressing the potential harmful effects of AI and suggests that the focus should shift to understanding how AI can help individuals become more engaged and effective members of society.

Keywords

💡Artificial Cognition

Artificial cognition refers to the simulation of human cognitive processes by artificial systems, such as AI. In the video, it is discussed as intersecting with human cognition in knowledge processes. The speaker, George Simons, is involved in researching this intersection, which is central to understanding how AI can be integrated into learning and educational systems to enhance student success.

💡AI in Education

AI in Education is the application of artificial intelligence technologies to enhance teaching, learning, and administrative processes in educational institutions. The video discusses the transformative potential of AI in universities, emphasizing the need for higher education to innovate and adapt to remain relevant in a rapidly changing world.

💡Adaptive Systems

Adaptive systems in the context of the video refer to AI technologies that adjust their behavior based on the needs of individual learners. They are highlighted as a key application of AI in education, aiming to provide personalized learning experiences that cater to the unique requirements of each student.

💡Metacognition

Metacognition is the awareness and understanding of one's own thought processes. In the script, it is mentioned in relation to how AI intersects with human cognition, particularly in areas of learning management and regulation, which are critical for understanding how AI can support complex problem-solving and learning processes.

💡Generative AI

Generative AI refers to AI models that can create new content, such as text, images, or music. The video discusses the growth of generative AI, especially after the emergence of models like Chat GPT, and its implications for various fields, including education.

💡Ethical Dynamics

Ethical dynamics pertain to the moral and ethical considerations surrounding the use of AI technologies. The video script raises concerns about ensuring that AI helps rather than harms individuals, emphasizing the importance of preserving integrity and security in the context of AI applications in education.

💡Personalized Learning

Personalized learning is an educational approach that tailors teaching methods and content to the individual needs and abilities of each student. The script discusses the benefits of AI in facilitating personalized learning, such as providing one-to-one relationships akin to the support offered by a tutor.

💡AI-First University

An AI-First University is a concept where AI is integrated into all aspects of the organization, from infrastructure to admissions, teaching, and research. The video presents this as a model for universities to adopt in order to innovate and adapt to the challenges and opportunities presented by AI.

💡Data Architecture

Data architecture in the context of the video refers to the infrastructure and systems that support the collection, storage, and management of data, which is crucial for AI applications. The script emphasizes the importance of getting data architecture right as a foundational step in effectively utilizing AI in educational settings.

💡Pedagogical Models

Pedagogical models are the methods and strategies used in teaching and learning. The video discusses the need to re-evaluate these models in light of AI, considering how AI can transform traditional educational practices and support new ways of engaging with learners.

💡Complex Problem Solving

Complex problem solving is the ability to address multifaceted issues that require a combination of skills and knowledge. The script mentions this in relation to the integration of human and artificial cognition, highlighting the importance of understanding how AI can augment human capabilities in tackling complex challenges.

Highlights

George Simons discusses the intersection of human and artificial cognition in knowledge processes and his role as a co-founder, chief scientist, and architect of SNH use human system.

Simons points out the misjudgment of higher education in its response to AI, highlighting the need for universities to be more responsive and capable in utilizing AI for student success and relevance.

He outlines four topic areas to be discussed, including literature on AI and learning, the intersection of human and machine, current trends in AI, and the implications for universities.

Simons emphasizes the importance of understanding AI not as a tool but as a node within our cognitive network, with significant educational implications.

Discusses the importance of re-evaluating human learning and knowledge growth with AI as a mediating and transforming agent in the ecosystem.

Presents research on the effects of AI in complex problem-solving and the critical components involved in the intersection of human and artificial cognition.

Simons talks about the rapid pace of AI decision-making and the inability for humans to intervene in real-time, necessitating a reevaluation of human cognitive functions.

He identifies the need for new theories of learning that integrate not just human-to-machine interactions but also machine-to-machine interactions in high-risk areas.

Simons reviews the practical applications of AI in classrooms, focusing on adaptive systems, personalization, profiling, prediction, assessment, and evaluation.

Discusses the benefits of AI in education, such as personalized learning and better administrative activity, while also acknowledging the challenges introduced by AI.

Ethical dynamics, curriculum development, and infrastructure questions are highlighted as significant challenges in the integration of AI in universities.

Simons emphasizes the importance of getting data and related architecture right as a critical challenge for universities initiating AI conversations.

He suggests that AI will impact every aspect of a university, from infrastructure to admissions, teaching, assessment, curriculum, and research.

Simons proposes the concept of an AI-first university, where AI is involved in all organizational aspects, and discusses six areas of priorities for such universities.

The importance of multi-institution collaboration and sharing data across operational and data science platforms is stressed for a university AI response.

Simons concludes by emphasizing the shift in education from teaching knowledge to developing human beings and preparing them to engage with non-human forms of intelligence.

Transcripts

thing he is George cens um I think

they're getting ready with all the last

details before taking the stage so let

me briefly introduce him uh he

researches how human and artificial

cognition intersect in knowledge

processes he's also a co-founder a chief

scientist and architect of SNH use human

system that is an organization building

resources to to respond to systems

impact on AI on learning and also

Wellness I think we're ready here now

are you jge okay please come to the

state welcome him George

Simons thanks so much for joining

us uh good morning and uh appreciate the

opportunity to spend some time talking

about what I think is a significant

misjudgment on the part of higher

education over the last certainly

several years but likely going back well

over a decade and that is a somewhat

fatigued and even baguer response to AI

as a mechanism for changing and

innovating the university sector as a

whole so I'm going to talk through what

I think is happening and what I think we

need to do as universities to be more

responsive and more capable to utilize

AI as again a mechanism for ensuring our

students are successful but also

ensuring that universities continue to

remain relevant in a pretty quickly

changing world I'm going to talk about

four distinct topic areas the bulk of

the talk I'm going to look at some of

the literature around Ai and learning

and this is just to give you a bit of a

sense on what do we know from literature

that works well in learning and learning

related processes I'm going to build a

little bit on what Charles was just

talking about which is the intersection

between human and machine it's a

co-creation process not necessarily

antagonistic process process I'm going

to talk very briefly two slides worth

about AI specifically and I'm just going

to detail what it is that AI does and

what some of the current trends are that

we're seeing in AI I assume everyone in

the audience doesn't need the 500th what

is AI primer so I'm just going to talk

about what's happening right now

specifically around llms that have an

educational implication I'm going to

from there go a little bit about what

does this mean specifically from a

university lens and how universities

might change and then finally I'll

present sort of a six area of priorities

that universities need to pay attention

to if they want to start getting more

actively involved in the AI conversation

so to get

started we're at an interesting time in

history in that we've are sort of at the

tail end of an extended period of

emotional turmoil as a society uh we

have seen a significant increase in

escalation in areas of emotional need or

in areas of L loneliness and mental

health impacts are certainly growing not

only limited to the effects of uh the

pandemic but just stats and indicators

prior to the pandemic that said hey as

people were not doing okay emotionally

and mentally some of the systems that

Society has created for us aren't

serving all of us equitably and that's a

significant Challenge and so there's

ways that we need to be better in how we

support and engage with Society writ

large not just with individual learners

but what happens is each time we have a

new technology we introduce a bit of a

spacing effect and that spacing effect

means social media as an illustration

initially came on and it allowed us to

connect with people from around the

world but nowadays that connection is

actually producing disconnection and so

what initially gave us the opportunity

to do new things with new groups of

people suddenly became become at odds

and in conflict with new groups of

people so the way social media has been

deployed by itself was naive and

effective but once you make it available

for algorithmic Distortion and for

propaganda suddenly it becomes harmful

and actually disruptive to the system as

a whole and so we need to keep that in

the back of our minds because the

lessons of social media on mental health

and on society Wellness will be almost

insignificant can compare to the threat

and the risk that AI will pose into the

public conversational sphere so each new

wave of Technology forces us to evaluate

the spaces that we occupy and how we

remain human in those environments and

that's one of the reasons I particularly

appreciate the uh the theme of the

education Forum here around that human

component in AI settings so when you

look at traditional learning literature

there's been a long period of

acknowledging that thinking and learning

doesn't just happen in our brains right

there's a range of theorists that from

embodied cognition to distributed

cognition to some externalization of

Concepts and ideas we're constantly

putting human knowledge into physical

things or objects or concepts in the

world and most established theorists and

philosophers would argue that you are

intelligent as a function of the

networks that you exist within and those

networks traditionally have been tools

and resources we've created such as

books and related artifacts but

increasingly now they're starting to

become systems that are AI compliant or

AI enabled so when I think of artificial

intelligence to me it's not a tool it's

not a resource that we use it is a node

within our cognitive Network and that

has significant implications

educationally

and that's because as a species we don't

exist in these systems as isolated

entities the best way to describe it is

we and not just as humans but species

all of life all of society coexists and

exists fundamentally as a function of

networks the idea of individual is

actually antithetical in terms of growth

opportunities and the advancement of

society all of our capabilities are a

byproduct of how we're Network and

connected so we did a paper a while ago

where we wanted to understand if we

bring AI into these learning processes

such as complex problem solving what are

the effects of that you know what are

the critical components that are

involved assuming that you agree with me

that networks are the foundational

underpinnings and so we looked at

essentially when you have human and

artificial cognition intersecting in

areas of metacognition such as

regulation and learning management in

affect related to things such as emotion

and trust and confidence and the way

that we connect with one another with a

sense of security and confidence what

does that look like or if you then take

and look at the cognitive practices

things like remembering what's the

importance of memory when AI is at your

fingertips or which parts of memory

remain relevant when AI is at your

fingertips because one of the things

that AI does in this conversation uh is

move capability questions to to a new

plane it's not that it makes those

things irrelevant it means that we are

related to some of those Core Concepts

differently than we perhaps have been in

the past and similarly with social

practices and collaboration and

engagement and working together so when

you bring AI into this process one

argument that I've been making to

colleagues for years is that every

single thing that we know and understand

about human learning and human knowledge

growth needs to be re-evaluated

with an understanding of AI as a

potential mediating and transforming

agent within that

ecosystem and so we looked at if you

take these two pieces and you bring them

together because that's what we

essentially see happening it's not that

we're saying AI is a tool off to the

side I'm arguing that AI is the first

injection of intelligence in the human

System since our neocortex came on line

so it is an alien intelligence it's not

exactly like us but it does certain

things that can make some stuff easier

for all its criticisms for its

hallucinations for its biases AI is a

type of an intelligence that we can

co-thinkers a period of these little

blips of sudden crashes uh there's a

research uh report that was put out by

Johnson where he said these systems

where AI is starting to make decisions

they're moving so fast that we are at a

point where there is an inability for

humans to intervene in real time meaning

it's machines have taken over large

swads of those kinds of processes and

what it's done for us we can't

participate in real time so the human

cognitive function is to escalate which

means we move to a higher plane because

we can't do the granual level

performance at the same level that AI

meaningfully can and that's produced

work such as as this paper by rwan and

all where they said we need to start

thinking about theories of learning that

don't just integrate human to machine

interactions it's machino machine

interactions that we need to think about

because there are sads of decisions in

some high-risk areas including medical

and Military where AI is making

decisions often without a human input

layer uh brought in and so to start

thinking about complex problem solving

and the integration of human and

artificial cognition into this kind of a

landscape is critical so a paper we did

a few years ago we looked at exactly

this question is what happens when you

have two types of intelligence that

maybe don't quite understand each other

but we know that meaningful integration

between the two is going to be important

for solving all the problems that

Humanity faces from homelessness to

inequality to climate change um how do

we begin to make those two play together

and what is that intersecting space

where learning and sense making and

meaning making happen meaningfully at

that level so we did a paper um in just

last year actually where we looked at

the literature that to date has looked

at Ai and uh its impact on the education

setting specifically what are people

doing with AI in classrooms in a

practical way not in a high flut and

future way that says oh we'll all have a

personal agent and we'll all be happy

and have a robot in our home but in a

practical way what's actually happening

in classrooms and so the number one set

of applications are ones that still

remain prominent which is adaptive

systems and

personalization that's been a holy grail

of education for decades but it says

rather than one student or one teacher

teaching 30 students everyone has a one

toone relationship like was mentioned

previously this is the idea of blooms 2

Sigma where the inclusion of a tutor can

move a c student to an a student with

the right level of support and guidance

profiling and prediction was an

important one that came up as well a big

part of what universities haven't done

historically is to understand their

students you know what are their skill

sets what are their capabilities outside

of a grade and so it's this idea of how

can we better profile and then if we

profile predict which students will

succeed which students are at risk of

potentially dropping out assessment and

evaluation is another important one and

then interestingly uh tutors were right

at the bottom at least of this cluster

it wasn't a huge area of use this data

would obviously be very different if we

were to do this report again in a year's

time because one of the top adaptations

of the a growth of gener of AI has been

tutoring and adaptive systems of that

type so the benefits then are

straightforward personalized learning

positive influence on the education

process um better administrative

activity from a university level as well

helping get insight into how students

are learning and then also as a way of

doing more effective assessment but that

doesn't mean everything is all

delightful because there's some

significant challenges that are

introduced with AI in this landscape one

probably top Remains the ethical Dynamic

how do we ensure that AI helps not harms

people how do we preserve the Integrity

how do we preserve the uh the security

of the student in this area of growing

Automation and increased technology a

lot of attention to curriculum devel

velopment how do we use AI well to

create courses and then a range of in

infrastructure questions that I'll

address uh once once I get a little

further toward the end the big research

gaps in the literature um are what you

would expect ethics keeps coming up top

of the list because that remains one of

the bigger unspoken challenges in the

University sector as a whole and not

just University across all of society a

lot more questions about methodology uh

this is a conversation was having with

uh my wife on this as well recently

which is in education we've typically

done we take a concept and we develop a

theory around it Theory sometimes is the

byproduct of extensive research and then

we use that to guide and shape decisions

going forward but now we're at a

slightly different landscape in that we

can use large swads of data and rapidly

move that forward to try and gain

insight into students and student

performance when we started to look at

this more from an llm side there was a

acceleration on a number of fronts but

the same questions remain profiling

prediction feedback remained key

concerns uh in the educational landscape

whether we're looking traditional AI or

emerging

AI we did a paper uh actually I think it

was this year um where one of the

outcomes was we looked at student focus

and student engagement when you bring AI

into the classroom setting and the

interesting thing we found was that

students don't necessarily learn from AI

they instead rely on AI which is an

interesting distinction uh it doesn't

have the same learning capability in all

settings as always it's a function of

pedagogical approach and pedagogical

models one of the big papers though that

I always refer to and this is an

important concept when we talk

methodology is that a lot of the

activity that happens in a classroom is

based on uh that happens in research is

based on a setting that's disconnected

from reality and an Brown did a

fantastic IC paper uh you know was it 40

plus years ago where she looked at this

design experiment that the entirety of a

classroom is a learning ecosystem for

learning research rather than these

oneoff experimental design settings and

that's exactly the kind of activity that

we try to do in digital spaces now

through the use of data and data

collection which we get from a range of

sources student Information Systems uh

instruments or survey instruments we

deploy Learning Management Systems we

can get a fairly holistic assessment or

lens of what a student is doing and

where she is in her overall learning

process so with that as a backdrop I'll

take the last 10 minutes to talk through

these final sections so if we look at

technology over the last few decades we

can say the open education movement

fundamentally taught us that we can

scale content with minimal cost

additions each new duplication of a web

page or a PDF is really insignificant

compared to the cost of duplicating a

new textbook a second thing that we

learn through open online courses or

mukes in some cases is that we can scale

teaching we can have a 100,000 or

500,000 students take a course and it's

much less expensive from a lecture lens

if that's the primary pedagogy in that

kind of a setting an AI is at the early

cusp I believe of teaching us that we

can accelerate and scale interaction so

the connections that we have on sort of

a onetoone basis from a tutoring

perspective the significant Trends I

want to identify here though relate to

where is the current state of AI after

chat GPT and the growth of generative AI

the hype that we had in 2022 and early

last year we're starting to see some

very practical groundings of these

Technologies not least of which is the

prevalence of AI in everything from our

cars to software to the platforms we use

growing multimedia and multimodal and

also a lot of atten being paid to open

source llms or open source software a

lot of that's driven by meta

interestingly enough and a growing group

of uh organizations notably stuff like

misil and others that are really

promoting open llms there's also

attention being paid to very small llms

which you're going to see more and more

on your Android or on your iPhone

devices uh fe2 fe3 actually just came

out uh at the end of April as well so

we're starting to see them accelerating

similarly AI pairings meaning AI with

traditional Robotics are starting to

come together and I think most of us in

this room will have a an aid driven

robot in our homes within the next 5

years doing routine related house tasks

a lot of attention now this is maybe a

little more relevant to some of you who

are running technical teams there's been

a significant acceleration of platform

technologies that make AI development

easy if you were to do something with an

llm 16 months ago or 12 months ago you

needed a fairly High technical

capability but now in environments like

AWS or vertex you can quickly run up a

series of models test and deploy uh with

a team of one who has fairly fundamental

understandings of the process um we're

also seeing a lot of I'll skip that one

uh more and more wearables wearable

devices uh rayb bands is an interesting

one again meta driving uh which is the

ability to have your glasses as you're

walking see a scen in front of you you

in llama 3 which is Meadows open llm uh

you can ask it what am I looking at

what's this picture and it will search

and provide an answer back to you uh

audibly on your on your uh glasses as

well and then a lot of tooling things

which is a little Beyond where we are

today but tools like dspi and Lang chain

that make this process of managing

multiple llm Integrations much more

effective so what are some of the inte

inte implications of this well first of

all I think AI will impact roughly

everything that University does there's

no sector that's not going to be

challenged by it and I do think it

represents a systems level challenge for

the sector and I don't think

universities see that and I don't think

many of them are responding as urgently

as they should because if you look at

one of the main things we do is we

generate knowledge and we communicate

knowledge that's our role as a

university and AI plays in all of those

territories I mean here's just a range

of tools that are knowledge adjacent

generating Technologies some of them

have been deprecated you know like

Galactica was briefly put out but then

paused but there's a lot that you can do

with this growing Suite of AI tools that

intersect with human creativity and

human knowledge

capability the system itself as an

Enterprise is already in a process of

unbundling it's no longer a

self-contained system a lot of what we

offer is increasingly being done by a

range of providers and we're going to

start to see exactly the same effect

happening in AI tools if you're a leader

in a university you're going to get a

range of providers and technology

companies coming up to you selling you

AI technologies that do everything from

uh tutoring to content creation to

assessment to student recruiting to

chatbot engagement and so on so it's a

constant influx of new technologies and

new approaches and so the way that we're

going to adopt as a sector is really

going to be one of three a direct

response to a simple problem a platform

based response or as a transformational

angle or transformational opportunity

from a system preserving lens the first

or the second one it's about just taking

Ai and helping it solve a problem like

advising or providing better student

support or the idea of a learner

co-pilot you know Microsoft co-pilot and

others are already making that available

or doing things like adaptive feedback

um that's what universities such as ASU

and what you're seeing with University

of Florida they're taking this kind of

an approach where they're largely going

out and just finding a problem and

solving it with some function of AI if

you want system changing approaches

though you need to start thinking very

differently about your literacies about

developing personal learning graphs and

personal models of a learner that

transcends a course even transcends

their lives computed curriculum not

pre-structured textbooks but curriculum

that's generated based on what a learner

knows and integration of Labor Market

needs into that educational process as

well well so we're talking about not

doing education as usual but doing

education a fundamentally different

way so the idea then is this

articulation of an AI first University

and an AI first University is one where

AI is involved in all aspects of the

organization from the infrastructure

through to admissions teaching

assessment curriculum and the research

process and I'll run through six of

those very quickly but you know one is

the infrastructure the pipeline the data

leg so any AI employment is

fundamentally a data challenge secondly

it's about building institutional

capability with AI like do does the

organization know what AI is and how AI

performs and what it does um thirdly

there's a range of questions that relate

to the leadership and policy and

governance how does the University

enable AI experimentation how does it

protect University reputation through

effective AI

engagement adaptive and responsive

teaching methods as noted this is

already prevalent in the literature but

how do we begin to use AI in such a way

that it is focused onetoone support for

Learners how do we improve the

personalized experience so that each

individual is met at her needs not just

cognitively but metacognitively

affectively socially and so on so it's a

very nuanced uh response to individual

learner needs and then also the

acceleration of research through the

utilization of AI uh doing a simple

literature review uh is now dramatically

different through the inclusion of tools

like elicit consensus or Iris

AI so I think at the end of the the

final several slides one of the critical

challenges I want to emphasize for

anyone that's initiating the AI

conversation is get the data and the

related architecture right more than

almost anything else this is a critical

challenge there are needs of building

capabilities institutionally what I mean

by that is being AI capable as an

organization and having the technical

capacity to train fine-tune models build

your own Bots those are expected but at

the in institution-wide concern of

infrastructure is critical um we just

released a paper for discussion uh

yesterday actually on a global data

Consortium where we tried to lay out how

should Universities at scale begin

collaborating and sharing data so that

you can learn from your peers rather

than try and do everything on your own

so the university AI response should be

through multi-institution collaboration

and sharing across operational data

analysis data science planes and then

ultimately addressing and driving impact

so it's a critical outcome uh we have a

white paper that's now out for review uh

from the American Council on education

if anyone's interested on that um final

points we're really getting at this idea

where most of education has been about

teaching people knowledge related things

you know the epistemology question and I

think we're now moving to the on ology

question like who are we as human beings

how do we develop human beings how do we

help people become more engaged more

productive and more effective members of

society you know any of the kinds of

things that are here like what is it

that we should be teaching how should we

be teaching people and Learners places

of being in the world how should we be

driving their capability to navigate

complexity to engage with non-human

forms of intelligence and then as a

byproduct of that to be sort of

proactive engaged and anticipating

potential harmful effects of AI as we go

forward thank

[Applause]

you thank you very much George wonderful

intervention yes for the next guest

thank you thank you very much again um

now we are moving forward H and I'm

going to switch switch again into

Spanish

much gracias

[Music]

ra

comp for

ET

for

for for

[Music]

spe

for

spe

Q

Sol

WhatsApp Instagram teams

can

for for

Sal

[Applause]

ra

for for

Ai and critical

thinking

for for

3:30 PM please we'll be back here thank

you very much for everything

[Applause]