Marco Cappato & Francesco Vecchi discuss how AI can empower democracy
Ep. 33

Marco Cappato & Francesco Vecchi discuss how AI can empower democracy

Episode description

Marco Cappato and Francesco Vecchi from the pan-European Eumans movement engage for a more transnational democracy and a more long-term orientation. In this episode we talked about the Civic AI initiative as part of the Eumans movement.

Download transcript (.srt)
0:00

Welcome on another episode of Democracy

0:02

Innovator podcast and our guest of today

0:04

are Marco Kapat and Franchesco

0:07

and

0:09

thank you for your time.

0:11

>> Thank you.

0:12

>> And you're working on um a humans,

0:17

right? What is a movement?

0:22

>> Humans is a panuropean movement of

0:25

popular initiative. What does it mean?

0:29

um a movement that is directly European,

0:34

so open to the membership of people all

0:37

around Europe um without the need of

0:41

passing through national organization.

0:45

So is a directly citizens movement um

0:50

directly at the European dimension and

0:54

is political but non-electural so

0:57

popular initiative because

1:00

we uh do politics in political

1:04

initiatives but through the tools of

1:07

civic particip participation at European

1:10

level for example the European citizens

1:14

initiatives

1:15

but also national, regional and local uh

1:18

level because we think that the two main

1:24

limits and problems of democracy

1:27

nowadays are first of all democracy is

1:30

mainly national but the problem of our

1:34

era are mainly transnational supra and

1:38

and infraational also and we also think

1:42

that uh uh elections, electoral

1:46

democracy is uh too uh concentrated

1:51

focused on shortterm consensus

1:56

um

1:58

in front of problems that need long-term

2:02

solutions. So there is the need to

2:05

complement electoral democracy,

2:08

national electoral democracy through a

2:12

robust dose, a strong dose of panuropean

2:19

citizen

2:20

u democracy, direct participation.

2:24

>> Yeah, let let me add a couple of things.

2:26

I do think that what Marco said is uh

2:28

really true. Um the point is that humans

2:31

was also born to fill in the gaps of a

2:33

few dissatisfactions.

2:35

um we believe of citizens in general but

2:37

especially know we started from the

2:38

Italian use case and the satisfaction

2:40

were not only the scope of the

2:42

participation meaning you know national

2:45

versus international but also the

2:47

satisfaction that a lot of widespread uh

2:51

a lot of campaigns a lot of ideas that

2:52

would be widespread ac among the public

2:54

so I don't know for example euthanasia

2:56

rights or abortion rights and other

2:59

things didn't really find any match into

3:02

the political debate because any party

3:05

is a bit worried of losing some specific

3:07

constituencies, they would be against

3:09

that. And since party politics is about

3:11

gaining consensus and gaining, let's

3:12

say, quantitative consensus, no one can

3:15

um can really afford losing that kind of

3:19

political support. Therefore, we do

3:21

believe that there are some topics that

3:23

should be discussed outside uh or at

3:25

least that whose discussion should be

3:27

complemented outside of a usual

3:29

representative democracy. This is the

3:30

case again for uh transversal political

3:34

uh initiatives.

3:37

>> And now you are both working on a

3:39

campaign called Civic AI, right?

3:44

>> Yes, we do.

3:46

>> What is CV AI?

3:49

>> H well uh this is a very tough question.

3:52

Uh maybe I can step in here. uh the

3:55

problem in defining civic artificial

3:57

intelligence is that we are first of all

3:59

speaking of uh a moving target under so

4:02

many different layers. The technology

4:04

itself is a moving target. The

4:06

technology itself is extremely complex

4:08

and you know it is multi-layered. It

4:10

starts from the infrastructure. Then it

4:12

gets to the layer of how data are

4:13

collect, stored, gathered and used. And

4:15

then there is the um then there is the

4:18

level of the algorithm and then there is

4:19

the level of the software and then there

4:21

is the level of the interface with users

4:23

and effects that artificial intelligence

4:25

might have on specific users and

4:27

citizens. Therefore u it's not only the

4:30

technology uh which is a moving target

4:32

but the regulation of the technology is

4:34

a moving target on its own. I mean we

4:36

have the example of the European Union

4:38

and other places in the world are

4:40

starting to uh regulating AI of course

4:42

with completely different approaches uh

4:44

compared to the European Union's one. Uh

4:47

but then another problem is also that uh

4:49

that I can say the dissemination of this

4:51

technology into the um contemporary

4:55

public sphere and especially the digital

4:56

sphere is constantly changing because we

4:58

are also getting used to the different

5:00

applications. So um is civic AI a fight

5:04

against disinformation generative AI le

5:06

disinformation is civic AI uh how can I

5:10

say um the campaign for ensuring that

5:13

each specific layer that I mentioned

5:15

earlier uh complies with European values

5:18

and democratic values? Is it a campaign

5:19

for increasing and enhancing civic

5:22

participation online also through the

5:24

support of CVKI chat bots or virtual

5:26

assistants or platforms? I mean the

5:28

point is that CVKI is a bit of

5:30

everything. Um maybe I I'll leave the

5:34

floor to Marco to explain what was the

5:36

original idea of CVKI and then I can get

5:38

back the word to explain you know what

5:40

we did in the last few months and what

5:42

we are uh trying to what we uh what we

5:45

plan to do in the coming uh weeks and

5:47

months.

5:48

>> Yes. uh well I will start from the

5:50

consideration that uh uh almost I think

5:54

99.999%

5:58

of investment

6:00

of huge immense amount of money poured

6:04

into uh artificial intelligence in the

6:08

current times is driven by either

6:14

commercial purpose profit company either

6:19

uh control, warfare,

6:22

uh security. Um

6:25

but there is virtually no investment or

6:30

really marginal

6:32

um

6:35

uh on the use of AI to empower the

6:39

citizen.

6:41

Um that's exactly the meaning of civic

6:46

AI artificial intelligence as a public

6:50

service. There is no major um invention

6:56

and uh technological advancement that

7:01

didn't have both the private uh

7:06

commercial use but at the same time the

7:08

issue of how this invention could be

7:13

provided as a public service. Let's

7:15

think about I don't know power,

7:18

electricity, gas, mobility, cars and so

7:21

on. Um the revolution that we are

7:26

experiencing that is already triggered

7:30

um is there is a lot of debate on how we

7:34

should regulate and limits the the risk

7:39

of abuse, manipulation and so on. But

7:42

there is no real discussion and most

7:46

most important no real big plan and

7:50

investment on how this revolution could

7:54

also not as an alternative but also be

7:58

used to benefit the citizens and empower

8:02

the citizens on those issues without

8:05

direct commercial interest. The first

8:09

example, the main one probably is

8:12

democracy.

8:13

Uh we know that democracy is uh in a big

8:18

is is in a big crisis. There is lower

8:22

and lower participation rates,

8:25

attendance rate to elections and people

8:29

are losing faith and trust on how

8:32

democracy could really help in

8:38

giving solutions to the problem of daily

8:41

life. So the idea of empowering through

8:46

artificial intelligence the possibility

8:49

of citizens uh of dialoguing and acting

8:54

with institution and among themselves to

8:58

find solution for their life for the

9:01

quality of their life for the quality of

9:04

the envir of the environment. This is

9:07

something that we cannot hope that will

9:10

be solved by big tech in Silicon Valley

9:13

or in China. This is a political problem

9:17

that need to be um replied with a

9:22

political solution at the public

9:25

institutional level. And the European

9:27

Union is the only dimension that could

9:30

in a way uh federate aggregate what is

9:34

already existing as civic use of

9:37

artificial intelligence to make them

9:39

interoperable to unite them and to

9:43

create a digital ecosystem

9:47

um to empower the citizens in the daily

9:50

life. So uh the this is the idea of

9:55

having of having also this path for the

9:59

future of artificial intelligence not

10:02

only the current paths that uh billions

10:07

of billions of billions of billions are

10:10

invested and attracted and attracted to.

10:14

Yes, if I may add something on this and

10:16

if I thank you Marco for mentioning you

10:18

know the need for creating a I can say

10:20

an ecosystem a digital ecosystem. This

10:22

is exactly uh one of the most important

10:25

things to clarify when we speak about

10:26

civic AI. Uh what we do have in mind is

10:29

trying to outline a European digital

10:32

public sphere. Uh since you know

10:34

technologies in general but especially

10:37

AI are influencing every single aspect

10:40

of how we do interact publicly. So

10:42

they're influencing the way information

10:44

is shared. They're influencing the way

10:46

we interact with public organizations or

10:49

public institutions. They are

10:50

influencing also our daily lives. We

10:53

must ensure that everything in our you

10:55

know in the in this mixed reality in

10:58

this hybrid reality that we are living

10:59

through you know which is mixing um can

11:02

I say physical assets and digital assets

11:05

we must ensure that this also exists at

11:08

the same uh this is per this is

11:10

permeated by democratic values at the

11:12

same time while this was the gen the

11:15

original idea of the petition maybe not

11:18

all of you know that any citizen in the

11:20

European Union can um how can I say uh

11:24

can submit a petition to the European

11:26

Parliament. Uh and this is was exactly

11:29

what Marco did supported by some experts

11:31

especially from Italy but not only from

11:33

Italy. Um the uh what happened to the

11:37

petition was that the petition was first

11:39

examined by the um committee of the

11:41

European Parliament dedicated to

11:43

petitions and then we were asked to

11:45

present it uh in April 2025 in front of

11:49

the of the committee. In that case,

11:51

Marco asked me to go simply because I

11:52

was in Brussels and um and after in that

11:56

case we gathered the support of all

11:58

representatives from all me or all

12:01

groups of the parliament for our

12:03

initiative. In that I must admit that in

12:05

that case we didn't not only uh present

12:08

again the petition was already existing

12:09

and you can still find on our website

12:11

but what we tried to outline is exactly

12:14

this the panuropean digital public

12:16

sphere and this public sphere was

12:18

composed of four main strands. The first

12:20

one is a civic assistance. So we you

12:23

know in our effort to explain what we

12:26

mean with civic AI the first uh

12:28

visualization that we got was a sort of

12:30

chatbot of victory assistant that would

12:32

help any citizen navigate not only the

12:34

complex framework of the European Union

12:36

but especially the all the participatory

12:39

opportunities the deliberative

12:41

opportunities they can jump in and in

12:43

general gather u insights and feedback

12:45

on how they can actively participate in

12:48

the European democracy. There are plenty

12:49

of of things that citizens can do. For

12:52

example, um provide feedback to open

12:54

consultations of the European

12:55

Commission. This is going to be

12:56

important also for later on. Um they can

13:00

on the have your say portal. They can

13:02

for instance uh participate in the

13:04

citizens engagement forum. Sometimes

13:06

there are uh participatory and

13:07

deliberative initiatives such as uh was

13:10

uh the conference on the future of

13:12

Europe. But in most cases people simply

13:14

don't know. And what we have to um what

13:17

we have to accept uh when you do

13:19

politics is that citizens are getting

13:21

used not even at having you know

13:23

services delivered as a softwares or as

13:26

as platforms but they're getting used of

13:28

having services delivered through

13:29

virtual assistants and chatbots. This is

13:32

the main trend that we see. So we would

13:34

like that public services not only as

13:37

the public the services uh given by

13:40

public authorities but also democracy

13:42

which is a public service in its own. I

13:44

mean it is so trivial admitting this but

13:47

still you know sometimes we do forget

13:48

that democracy should be a public

13:50

service in its own well we must ensure

13:52

that they are provided in the most

13:54

citizenfriendly way uh why am I

13:57

mentioning all of this was just in fact

14:00

the first trend that there are other

14:02

trend for example we mentioned the

14:04

participatory dashboard so let's say a

14:07

sort of platform that once you sign in

14:09

with your ideally European digital

14:11

identity according to the data you

14:13

provide to the specific platform. You

14:15

see all the participatory initiatives

14:17

that you can uh partic that you can join

14:20

depending of course on your res

14:22

residency rights, your nationality

14:24

rights, your citizenship rights and and

14:26

so on and so forth. And third there was

14:29

the liberation. creating a sort of tool

14:32

that would help public deliberation

14:34

through all the techn including AI but

14:36

through other technologies that help

14:39

create collective intelligence among

14:41

topics and I mean these technologies

14:43

already existed again in the platform

14:45

for the conference on the future of

14:46

Europe. So we're not saying to deploy

14:48

something completely new but just to

14:49

adapt what's existing to a slightly

14:52

different scope. And finally but this is

14:54

probably one of the most important

14:56

layers um there is what we call the

14:59

democracy data commons. So the concept

15:01

of data commerce and I know that Marco

15:03

probably would like to add something on

15:04

that as well. Um but the the concept of

15:07

data commerce is creating these sort of

15:08

data spaces that are public not

15:11

necessarily publicly managed but

15:13

definitely publicly owned and publicly

15:15

controlled and governed not necessarily

15:18

again by a government like let's say the

15:20

European Commission but it might be by

15:22

third authorities or independent

15:24

authorities like the European data uh

15:26

protection supervisor but this data uh

15:29

this data space would collect all the

15:30

data regarding the democratic

15:33

participation online and civic

15:34

participation online. Why is it

15:36

paramount? Because currently there are

15:38

plenty of participatory tools, participy

15:42

tools, but sometimes they are either um

15:47

can I say the data is either stored by

15:49

specific private oriented actors or they

15:53

they are stored even outside of the EU.

15:54

There are some platforms that are based

15:56

in the UK and this of course would

15:59

create some issues in for any um can I

16:02

say panuropean digital democracy tool

16:04

because I mean if your data is not even

16:06

under European law then of course you

16:08

know this raises issues. I could go on

16:10

with the next step. So probably uh I

16:13

would like Marco to spend a couple of

16:14

words on data com because he recently

16:17

had an idea also on this. I would like

16:20

to know the Alexandra's reaction because

16:22

we gave you a lot of information. So

16:27

what what what idea have you made out of

16:30

all of this of what civic artificial

16:33

intelligence could be? Sorry if I ask a

16:36

question instead of replying.

16:39

>> I like the question and I'm thinking

16:41

about all the interesting things that

16:42

you mentioned mentioned earlier. Um I'm

16:46

thinking about how information is power.

16:48

I mean if citizen knows what they can do

16:52

then they have also the power to do it

16:54

but uh as Franchesca said like often the

16:56

citizen doesn't know what he can

16:59

actually do and uh

17:02

and I was wondering like uh this uh

17:05

dissatisfaction that citizen has so we

17:08

have seen also how people maybe uh I

17:11

mean they still vote there there are

17:14

election but we cannot say that

17:16

representative democracy is at

17:19

uh the itest um it's not the itest

17:22

moment for representative democracy

17:25

and uh I totally agree about uh uh the

17:29

possibilities that we have now using

17:31

technology

17:32

uh to make um

17:35

um politics more accessible for citizens

17:40

uh because as you said like um

17:42

Franchesco with a chatbot maybe it's

17:44

easier to uh to tell to a citizen and

17:48

what he can do or what are the

17:50

possibilities

17:52

and uh I'm thinking also about the the

17:55

digital ecosystem that is also something

17:57

very important

17:59

uh because probably uh I can imagine

18:02

that several platforms are needed

18:04

several expertises

18:06

and um

18:08

and yeah I totally also agree about uh

18:12

how to that uh it it is necessary to

18:16

break the circle

18:18

of um of technology used for control or

18:22

for commercial reasons. Um

18:26

and actually uh it was uh yesterday I

18:29

was at the table with some uh PhD

18:32

students that that are studying these

18:34

kind of things and uh we were also

18:36

thinking about what is the CVKI uh

18:39

because I asked them uh the question and

18:42

uh it is actually not an easy uh thing

18:46

to

18:48

uh it's not actually easy um but yeah I

18:54

I think

18:54

>> maybe do some examples that could be

18:57

useful. Maybe also for for those who

19:01

probably could think, okay, but

19:05

do we really want the state or Europe or

19:09

the European Union be an actor in the

19:14

artificial intelligence race? Is this

19:18

realistic or is really something that is

19:21

suitable even more than realistic? Well,

19:24

um let's think about for example

19:28

platforms for civic participation

19:32

uh nowadays if you want to take an

19:36

initiative maybe in your own

19:38

neighborhood in your own city council uh

19:42

because you need to ask uh I don't know

19:45

the respect of a park of a green area

19:49

and so on what you do what are the tools

19:53

Usually you have this alternative. On

19:56

one side you have uh social media. You

20:01

can open pages uh on Facebook, launch

20:04

debates on Tik Tok and Instagram and so

20:07

on. But you know that the algorithm

20:11

will uh um will will in a way will push

20:18

um will push the the the content would

20:24

push the content that are polarized

20:29

uh and aggressive or with a very strong

20:32

idea and proposal. But civic

20:35

participation should follow another

20:38

logic because you don't need to sell

20:41

publicity advertisement products. So uh

20:47

the unfortunately

20:50

civic participation

20:52

uh the tools the non-commercial tools

20:55

for civic participation are boring or

21:00

non-existence or technologically

21:03

I mean not even digital sometimes you

21:07

don't even have a possibility of digital

21:09

interaction. So let's think about how

21:13

important would be to have a platform

21:16

for digital participation empowered with

21:20

artificial intelligence user friendly

21:23

and at the same time with an algorithm

21:26

conceived in a way uh that uh dialogue

21:31

is promoted

21:33

uh to speak but also to listen to have a

21:37

real interaction among citizens. Okay.

21:41

And uh many

21:44

local civic digital things exist around

21:47

Europe but they are not interconnected.

21:51

they are not uh uh you have you don't

21:55

have a single sign on

21:58

uh to uh like

22:02

to have access to all the digital tools

22:05

of interaction with the state and with

22:08

other citizens around Europe. So this

22:11

could be one of the concrete outcome of

22:17

an ecosystem of civic of European civic

22:20

AI to have uh interoperability

22:24

of platform of democratic

22:27

participations.

22:29

uh and this is not something that a

22:32

giant in Silicon Valley or in China

22:35

could do even because uh democracy is a

22:40

two delicate things. You need to have

22:44

for example when we are talking about

22:47

signing with a formal legal value a

22:51

petition or a proposal of referendum or

22:54

a ballot. the state is the only out is

22:58

is the only authority that can identify

23:02

and certify that signature as as as a

23:06

real one. So to uh to ensure the

23:10

legality of the process. So this is why

23:13

we think that even if

23:16

um the technological race toward

23:20

artificial intelligence is not something

23:23

that the state could play or should play

23:27

a competition role. Uh from the point of

23:32

view of the user of the citizens

23:35

uh the empowerment with artificial

23:38

intelligencebased system could be

23:40

crucial in having their rights,

23:44

democratic rights but even so social

23:47

rights on civic participation for public

23:51

goods for example for the as a user of

23:54

public goods. uh this dimension needs to

23:58

be uh empowered through a political

24:01

choice of investment by public

24:04

authorities.

24:07

>> Yes. Again um let me let me take a step

24:10

back to move a step forward. So um

24:14

whenever we speak about CVI sometimes

24:16

people get the impression that we do

24:18

want the European Union or whatever a

24:20

public actual in the regard to develop a

24:23

platform or develop a chatbot and put it

24:26

in the hands of the citizens. uh I mean

24:29

developing this kind of tools is

24:30

extremely expensive and not always

24:32

public institutions do have the talent

24:34

pool needed to develop something and

24:36

would effectively work and in that case

24:38

case again you know as Marco was saying

24:39

I mean we they would put themselves into

24:41

competition with the private actors that

24:43

have more money more talent and more

24:44

expertise in the field so why did I

24:48

mention to take a step back let's have a

24:50

look at the European digital wallet and

24:52

the EI does framework what did the

24:54

European do union do in in that regard

24:58

They first of all developed a framework

25:00

that would define would outline some

25:02

rules some guidelines on how to deploy

25:04

softwares and tools for uh trust service

25:07

providers for digital identity that

25:09

happened I mean I don't remember if five

25:11

10 years ago the first version and then

25:13

recently the second version was uh

25:15

published

25:16

according to the frameworks that were

25:18

deployed by the European Union basically

25:21

the assumption was okay if you comply

25:22

with these frameworks and you deploy a

25:24

specific tool then this tool will

25:26

receive that kind of public

25:28

certification that Marco was mentioning

25:30

earlier. They created uh a sort of uh

25:35

market incentive. In that case, for

25:37

example, in Italy, we had the deployment

25:39

of two main tools. On the one hand, you

25:41

have speed. On the other hand, you have

25:43

uh the electronic digital identity. Two

25:46

completely different structures. They

25:48

are slightly different, but let's say

25:49

that both comply with the eidas

25:52

framework. What happened then that with

25:55

that with those tools for example Monaco

25:57

and especially Luka Kushon and other

25:59

actors in Italy they both forgetting the

26:02

possibility of signing digitally

26:04

referendums which means you know that

26:07

these platforms that just received

26:09

market incentive they say because they

26:11

were complying with rules with laws so

26:13

they could be used for anything public

26:15

and official then now they are actually

26:18

used for signing official referendum or

26:21

at least the proposal sorry market and

26:23

make a mistake. Why am I mentioning all

26:25

of this? Well, because for example,

26:27

something that the European can you

26:29

union can do is this. First of all,

26:32

write a sort of framework for what is uh

26:35

what is needed for a electronic

26:38

participation tool to be compliant with

26:40

the EU law and to be officially recogni

26:43

for any interaction happening on this to

26:45

be officially recognized by the European

26:47

Union. Second, deploy as I was

26:49

mentioning a sort of data space for

26:51

democracy in order to and binding every

26:54

actor uh private or public may be to uh

26:58

store the data and to manage the data

27:01

through that data space which is public

27:03

publicly owned. Of course you know

27:04

people it is open but people cannot get

27:07

access to um personalized data but they

27:10

can only access to anonymized data for

27:12

doing research and whatever and then

27:15

deploy always the European Union a sort

27:17

of a single sign on platform may it be a

27:20

chatbot may it be a I don't know a

27:23

dashboard or whatever uh where you can

27:26

sign up with your European digital

27:28

identity and once you're signed up I

27:30

mean depending on your data okay you get

27:34

what is it that you see? What you see is

27:36

not only the participatory opportunities

27:38

that you're in, but every municipality,

27:40

every level of governance can upload a

27:43

participatory uh process online,

27:46

choosing themselves the specific tool

27:48

they use, but with the data that goes

27:50

again to the democracy data commerce.

27:53

Why is this solution more feasible than

27:55

deploying a chatbot uh by the eur the

27:58

public administration? Because for

27:59

example this would not only um how can I

28:03

say um this would uh only require

28:07

interoperability of all the systems

28:09

especially with the European digital

28:11

wallet. They would not this would not

28:13

necessarily require a centralized uh

28:16

data storage. This would not require a

28:18

centralized tool but it would in fact

28:21

allow for a proliferation of

28:22

decentralized tools.

28:24

So on decentralization in fact I can

28:27

move a step forward. You know you asked

28:29

us it's not so easy to understand what

28:30

civic AI is okay everything that I just

28:33

mentioned is not even AI I mean there

28:35

could be a virtual assistant or a

28:37

chatbot that would help you navigate

28:38

across all these platforms or whatever

28:40

there could be some machine learning

28:42

systems again you know to uh regulate

28:45

the debate online according to specific

28:47

rules there can be some um I don't know

28:50

um yeah some complex data analysis for

28:54

creating collective intelligence on

28:55

specific participatory processes again

28:58

these are also AI technologies but can

29:00

be achieved even with other

29:01

technologies. Um but when we speak about

29:04

civic AI we could also mention again for

29:07

example the infrastructure layer. So

29:09

currently as Marco was saying the AI

29:12

market is extremely centralized by few

29:15

monopolies that own not only can I say

29:18

that the software and they do not really

29:21

allow anyone to understand what are the

29:23

weights what is the whole process they

29:25

use because of course they base your

29:27

business model on that but the problem

29:29

is that often times what is centralized

29:31

is infrastructures

29:33

uh I've recently gone to the I recently

29:35

been to the Mozilla uh festival in

29:37

Barcelona and there I attended to a

29:39

session that was speaking of how a

29:40

federated AI system is possible. So

29:43

different organizations, different

29:45

places can have their I don't know data

29:47

servers or or whatever is need to deploy

29:50

an artificial intelligence system and

29:52

then what they can do is to federate the

29:54

system so that the data is still owned

29:56

by any specific organization but their

29:59

AI models are trained on all the data.

30:01

So they can run all over the data but

30:03

yet you know this data doesn't get

30:04

centralized. So again why am I

30:06

mentioning this when it comes to civic?

30:08

Because one of the questions that we

30:09

have to ask ourselves is do we want AI

30:12

to be held by a monopoly may it be

30:14

public or private or do we want it also

30:17

to reflect the uh values of democracy of

30:22

even competition in a way that are

30:24

embodied in the European Union at the

30:26

infrastructural level as well. And there

30:28

again I mean we can start conversations

30:29

on for example um even on blockchain

30:32

technologies on distributed ledger

30:34

technologies and how to use them. But

30:37

just to this is just to understand that

30:39

beyond visualizing what civic AI is.

30:42

What is important to understand is that

30:43

artificial intelligence is a general

30:46

technology as much as was the internet

30:48

and as such it affects many different

30:50

layers and uh translating our democratic

30:53

values and our civic values into AI to

30:56

make to enhance a civic participation

30:58

and not reduce it you know through

31:00

profilation and through um I can say the

31:05

increase of misinformation whatsoever.

31:07

Well, we really need to reflect at all

31:09

the different layers of this technology

31:11

and in this regard probably Europe if it

31:13

would if it would adopt a different

31:15

approach compared to the other two

31:17

giants especially US and China well

31:20

probably it would really find their way

31:21

of being competitive in the market

31:23

through a more democratic approach to

31:25

technology which is not only reflected

31:27

in I don't know huge words like

31:28

trustworthy citizencentric and blah blah

31:31

regulation that is only risk based as it

31:33

is now

31:36

It's very interesting the the data space

31:39

uh thing. It's also something that with

31:41

some people that are developing civic

31:43

tech tools we were thinking about uh uh

31:46

because let's say there is a sit citizen

31:49

assembly I don't know in Italy and then

31:50

another one in uh in Germany then uh if

31:54

the topic is similar then it will be uh

31:57

worth it maybe to have a connection

31:59

between the two uh citizen assembly to

32:03

understand better what citizens want and

32:07

uh

32:08

>> and yeah it's it's are interested in

32:10

these things.

32:11

>> And

32:13

maybe we could uh we could raise a

32:16

question which are the only data that uh

32:22

uh a dict a dictatorship will never have

32:26

access to because we know that for

32:29

example in China uh there is no

32:34

privacy issue. So every data are

32:38

collected

32:40

uh in a massive way without privacy

32:43

restrictions. So

32:46

we think we tend to think which is true

32:50

they have a huge competitive advantage

32:53

in terms of the amount of data they can

32:58

use to train artificial intelligence for

33:02

commercial and control purpose. But

33:05

which is the data set they will never

33:08

have the democratic participation data

33:12

set. If we as Europe, if we would make

33:17

systematic

33:19

the promotion of civic participation,

33:23

citizens assembly, local democracy,

33:26

local referendum, local ballot proposal,

33:30

regional, national, European and those

33:34

data anonymized of course how

33:40

huge value we could extract. act from it

33:43

in terms of what are the needs of the

33:46

people, the priorities for their own

33:49

lives. And if you combine this to the

33:53

data

33:55

that are owned by public authorities

33:58

because we also tend to think okay now

34:01

Silicon Valley they have all our

34:03

personal data because we are on uh the

34:06

iPhone and Facebook and Tik Tok and so

34:09

on. But uh we don't have to forget that

34:15

uh for the time being maybe in 10 years

34:17

it will be different but the huge the

34:20

the biggest amount of data it's still in

34:24

the hands of the state even if the state

34:28

is not using as uh Silicon Valley is

34:34

using but the health data for example

34:38

uh are detained by the state uh the

34:42

justice related the justice system data

34:46

but all the the public activity that are

34:49

producing data that sometimes are not

34:53

even collected by the state. So this is

34:57

why uh at the European level we could

35:00

create a dimension

35:03

for the use for the collection and use

35:07

of public data for public purpose. And

35:10

we could al we could also add the

35:14

possibility of voluntarily donate

35:18

our personal data anonymized of course

35:23

to uh to in a way to to to nourish to

35:28

make grow this uh public data data set.

35:32

So uh um I think that even at the level

35:36

of data that we tend to think that is a

35:39

lost battle because the giant of

35:42

artificial intelligence are treating

35:45

amount of data that the the public

35:49

entity will never will never be able to

35:52

use. I think this is not true

35:55

is it is not true. If we choose to

35:58

invest on the creation on the treatment

36:03

on the collection and the treatment of

36:05

those data, of course, if we stay uh

36:09

steady and looking what is happening

36:12

without doing anything, I think that uh

36:15

unfortunately the the the state and the

36:19

and the public dimension of institution

36:22

and the public interest will become less

36:26

and less That's relevant because uh if

36:30

the world is changing without and

36:34

outside public institution and public

36:38

interest and the democracy well

36:41

democracy will lose value. It will not

36:44

be any longer an activity

36:48

uh

36:50

considered

36:52

strongly considered by the people as an

36:54

activity able to improve their lives and

36:58

this would be the final defeat for

37:00

democracy not the crisis but but really

37:03

the defeat. So this is why to invest on

37:08

on this

37:09

macro level that we called civic AI and

37:12

as Francho explained can mean many many

37:16

different things is really a political

37:19

strategical choice that has to be made

37:23

but it's a eurent choice because in 10

37:26

years the amount of money that would be

37:29

collected for commercial purposes if we

37:33

don't do The same thing for public use

37:37

purposes would be

37:40

would be

37:43

so so bigger uh that the it would be a

37:48

lost battle to try to include public

37:52

interest logic and citizens le logic in

37:56

the race in the in the artificial

37:59

intelligence race.

38:03

Well yes this let me add something I

38:05

mean let me go back to what I said

38:06

earlier I think that in the last few

38:09

decades we forgot that democracy is a

38:11

public service I mean let's think about

38:13

that you know most of the public debate

38:16

happens on digital platforms that are

38:18

profitled and private owned not even

38:20

inside of the European Union uh I mean I

38:23

I think that both Mark and I have

38:25

nothing against businesses okay that's

38:27

not the problem but the problem is that

38:29

if the public debate happens mostly on

38:32

private own um platforms. It means that

38:35

those who uh own the platform can

38:38

dictate the rules and this is the reason

38:40

for which I mean back in the days I mean

38:43

let's not speak about Trump and now you

38:45

know what's happening for example on

38:47

platforms like meta Facebook and blah

38:49

blah uh for what concern for example

38:51

feminist content and whatsoever but

38:53

before Trump stepped in these platforms

38:56

that are mostly uh US based would follow

38:59

the the the policies of the the

39:03

government back in the days which was

39:04

the democratic government and therefore

39:06

I mean you saw that Donald Trump was

39:07

just banned from uh Twitter after

39:11

January the 6th and I mean there were

39:13

other uh contents that were you know

39:16

censored or limited and whatsoever. I

39:19

mean even though we can agree that those

39:21

content must be censored the point is

39:23

that who is in charge of censoring them

39:25

and this must be the public if we decide

39:28

to censor them or this debate must be

39:31

held on a public platforms until it is

39:34

held on private platform this is going

39:36

to be a problem but most of us simply

39:38

accepted that since there are these huge

39:40

marketplaces let's say this huge single

39:42

sign on prices so I mean one size fits

39:45

all and everyone has to go there if I

39:47

were the representative newspaper I

39:49

would seriously think but do we really

39:51

want to communicate what we do on this

39:54

social network uh and this is just one

39:57

one case okay uh let's take another one

40:02

uh there are some especially local uh

40:05

public institutions that are already

40:07

deploying chat bots and virtual

40:08

assistants to interact with the public

40:11

that often times they are extremely

40:13

limited like they they only answer to

40:16

you know questions about what are the

40:18

services that they provide. And if you

40:19

ask a questions for in question for

40:21

instance uh who is going to be the

40:24

candidate at the next municipal

40:26

elections in my town, they do not

40:28

provide that answer. They don't have it

40:30

even though they are the public

40:31

authority and at some point you know

40:32

they will have to register you know

40:34

these elections how the who will win or

40:36

not. Uh but the other problem is also

40:39

that um it's I think it's news from the

40:43

last month um Microsoft so meaning

40:47

Copilot

40:48

uh bought the license for a lot of data

40:51

from Harvard Medical School. Most likely

40:54

this happened because they want to make

40:57

copilot the most trustworthy uh chatbot

41:00

for what concerns data on uh even basic

41:04

health which means that you would ask

41:05

copilot if you can take paracetamol

41:07

while you're drinking or vice versa.

41:10

Okay. But these kind of basic medical

41:13

questions should be deployed by the

41:14

public. I mean this should be a question

41:16

of you know public uh health services.

41:19

Why is it a private uh company that

41:21

would get all these interaction data and

41:23

understand what are the needs of the

41:24

citizens and then can use the data to

41:26

provide even further services.

41:29

There are I mean again neither Mark or I

41:31

are against businesses. Uh personally I

41:34

have no issues with that but the point

41:36

is that there is a series of services

41:38

that must be deployed by the public or

41:40

otherwise we are creating a very tricky

41:42

situation.

41:45

>> I totally agree. I'm thinking that maybe

41:47

this uh difference is also something

41:50

that uh um I mean US and Europe have

41:56

different approach to health uh one that

42:00

is more public and one more private and

42:04

um

42:05

I have a lot of other questions but I

42:09

don't know if you have more time

42:12

um

42:13

or if

42:14

>> I can stay until one as I said Oh, I

42:16

don't know Mark if he has to step out

42:18

already.

42:19

>> I have a few minutes because I need to

42:21

go. So if uh there is a last question

42:24

for me then I will leave you for the

42:26

final. Yeah, it is. Uh so all this

42:30

ecosystem, all the data space and so on

42:34

could maybe lead to some uh uh to a

42:39

society where all citizens are also

42:42

active maybe uh where they can actually

42:46

participate. Do you feel that this is

42:48

possible? I I mean in in in society

42:53

already

42:54

um cooperation

42:57

has plays a big role. We are not only

43:02

uh

43:04

I mean there are plenty of uh nonprofit

43:09

dimensions of our private life uh even

43:13

uh from from family relationship and

43:17

friends and also the relationship with

43:21

the community and and we know how how

43:24

this is important as Franchesco said is

43:27

not against

43:29

uh uh the the the market or

43:33

u

43:35

is is not alternative. But we do a lot

43:38

of things because we like to do to to

43:43

for cooperation and uh

43:47

in a way to to be useful to the others

43:50

or also for our uh own interests through

43:55

democratic participation

43:57

because we have to defend our rights and

44:00

even our economic interests. So the the

44:04

political dimension

44:06

uh and the social dimensions are

44:11

a very important part of our society.

44:15

The risk is that if artificial

44:18

intelligence empowers

44:22

only the part of our life

44:26

uh which is

44:28

um dominated by the logic or is included

44:34

in the logic of uh market interaction.

44:39

the social dimension of our life risk to

44:43

lose

44:45

uh weight to lose importance in our own

44:48

life and at the contrary I think that uh

44:53

if we would be able in an easy way in a

44:57

consumerfriendly way to reinforce to

45:02

raise the part of our life dedicated to

45:05

society and the others we would do that.

45:09

So I think that well it would be naive

45:12

to to think that everybody want to

45:15

participate that everybody want to uh

45:19

make political initiative is not what I

45:22

think. But at the same time, if it would

45:26

make um if if it would if it would be

45:30

made much more easier to do so,

45:35

uh that would we would uh uh see a huge

45:40

increase in quantity and quality of

45:44

civic participation and socially driven

45:48

action. see this. I really do believe

45:50

because human being is a social. Sorry

45:53

if I interrupt you but I mean uh just to

45:55

ask you I mean what happened with the

45:56

Italian platform for uh presenting the

45:59

referend online

46:01

>> there was a blossoming of proposals was

46:03

>> this was a huge change because nowadays

46:07

you have uh a lot of ballot and even

46:12

referendum proposal that are made easier

46:15

at least in the collection of signature

46:18

phases. But of course uh democracy is

46:23

not about just about voting but not even

46:27

just about signing. Democracy is

46:30

something more is uh uh being formed

46:35

discussing is a dialogue. There are a

46:39

lot of processes that should be

46:42

facilitated before the moment of voting

46:46

or signing something. So this is the

46:49

ecosystem with that we need to build

46:52

thanks to digital innovation and we are

46:56

not doing so if not in in marginal ways

47:00

in marginal experiment at the local

47:02

level or sometime at national level but

47:05

there is no the um

47:10

the the the the political

47:13

uh really decision of following doing

47:17

this path of choosing this priority for

47:21

the evolution of the democratic state.

47:24

This would be important to provoke uh

47:28

with the with this action. This is what

47:30

we are asking to public authorities to

47:33

do.

47:37

>> Thank you a lot.

47:40

>> Thank you. I listen to you but I will

47:43

not speak anymore because I'm on the

47:45

move. Thank you. Thank you again.

47:50

And

47:50

>> yeah, I don't know if you have another

47:52

question because

47:52

>> yeah, maybe another question could be

47:54

like uh if you have a message uh for the

47:58

people that are actively searching for

48:00

new solution for

48:03

uh related to citizen participation,

48:05

maybe they're developing software. Maybe

48:07

they are trying to do something like

48:10

CVKI but I don't know in another

48:12

continent somewhere else like um yeah if

48:16

you have a message for them uh

48:18

>> yes I mean this is a message not only

48:19

for them but for ourselves as well. I

48:22

mean uh there are plenty of platforms

48:26

for civic engagement and democratic

48:28

participation online. There are plenty

48:30

solutions, believe me dozens. Uh, but I

48:33

think that most of them have already

48:34

been interviewed in this podcast in my

48:36

in my point of view. Um, sometimes when

48:39

we deploy these tools, we tend to think

48:42

more of the offer

48:45

and supply, let's say, and lessen the

48:48

demand. Uh, this is something that Mark

48:50

and I would like to investigate,

48:51

something that humans keeps, you know,

48:52

reflecting on. I mean, is there really

48:54

demand for more democracy? Is there

48:56

really demand for more participation?

48:58

And under what terms, in what shape?

49:01

What I was mentioning earlier, you know,

49:02

about that everyone is getting used to

49:04

interacting with chatbots and virtual

49:06

assistants as the first way of

49:07

interacting digitally. That was not the

49:09

case five years ago. That was not the

49:11

case 15 years ago. Um, we I do think

49:15

that we should not only think of the

49:17

tool and make the tool perfect, but

49:19

thinking of something that is really

49:21

used and effective for citizen and

49:24

answers to their demands. Otherwise we

49:27

are creating cathedral in the desert.

49:30

I totally agree about uh listening to

49:33

their demands because uh sometimes uh

49:36

this is a question that that I ask I

49:39

asked myself a lot of times why people

49:42

are not participating

49:44

and then uh I thought maybe they are

49:46

participating already. It's just that uh

49:49

I don't see it

49:51

and so I really agree that we should

49:53

listen and be

49:54

>> No, they are participating. I mean um

49:58

Marco said it but he didn't use these

50:00

words but p when we say panuropean we

50:04

mean transnational and translocal. Uh

50:06

something I were getting while

50:08

interacting with pe especially with

50:09

people that are not Italians even though

50:11

Italian base is our strongest uh

50:13

constituency is that they do engage a

50:16

lot but at the local level while

50:18

Italians are mostly used to engaging at

50:20

the national level or you know levels

50:22

that are not necessarily local. Uh why

50:24

is it relevant? Because I mean if you

50:26

deploy a tool you have to understand

50:27

that every different uh let's say at

50:30

least people at the if we if we are

50:34

thinking about the European Union let's

50:35

take the member state and let's define

50:37

the people the citizens of the member

50:38

states each people interacts with

50:40

democracy in completely different ways

50:42

and this is the result of centuries of

50:45

uh interactions at the local national

50:48

regional level and it highly depends on

50:49

the history of the specific country or

50:51

member state. um there's something to be

50:54

taken into account and the other point

50:56

is uh why have there been movements such

51:00

as um

51:02

um me too bright for future and then

51:06

there

51:08

black lives matter and so on and so

51:09

forth what or even now currently there

51:12

is my voice my choice which is a

51:14

European citizens initiative that

51:15

collected way more than 1 million

51:16

signatures and now it's interacting with

51:18

the European Parliament the European

51:19

Commission to push forward abortion

51:21

rights in Europe

51:22

uh why did these things work? Because

51:26

there was a strong demand behind there

51:29

were decades of debate or public opinion

51:32

built in this regard. I mean this is why

51:34

you know even participatory and

51:36

grassroots initiatives thrive the

51:38

regardless of the platform they use uh

51:42

but unless there is that demand people

51:44

are less involved in participating and

51:47

what is interesting in my point of view

51:49

is that I don't know if it is just

51:50

because of digital tools whatsoever but

51:52

clearly people are less engaged with

51:54

traditional politics that requires a

51:56

sort of tribal uh to accept you know a

52:00

series of values a series of proposals

52:02

you were part of the tribe and you

52:04

couldn't get out of the community. But

52:06

people can engage on single issues.

52:10

Even with completely different

52:11

perspectives, they can agree on a

52:12

specific topic. And this is the case for

52:14

example on Newton major rights. There

52:16

you have people from all over the

52:18

spectrum from the far left to the far

52:19

right depending on what do they do

52:21

believe is their need. They can you know

52:23

just forget that they come from a

52:25

tradition of from a political tradition

52:27

and then support an initiative. So I do

52:30

think that you know when it when it

52:31

comes to thinking about the needs of the

52:34

the demand of the service we don't have

52:36

just translate existing processes into

52:38

the digital world we do have to

52:40

completely rethink the processes

52:42

according to as Habermas would say this

52:46

new structural change of the public

52:47

sphere.

52:51

>> So let's change how this is uh

52:56

is actually working. I mean uh let's

52:59

think about new ways to make uh

53:01

participation easier u so that it can

53:05

happen everything that you and Marco

53:07

described in the

53:09

>> yes if you want to contribute to how we

53:11

think about how democra democratic

53:13

participation should happen digitally I

53:14

mean uh there is a page on our website

53:17

where you can join the campaign joining

53:19

the campaign means that you can get

53:21

access to our working groups and you can

53:23

co-create with us the proposal I mean

53:25

now we spoke about a lot we We gave you

53:27

a lot of information, a lot of ideas and

53:29

most of them are not Marcos nor minds.

53:32

They are just you know the ideas of the

53:34

community that got integrated in all our

53:36

activities. We try to co-create in this

53:38

campaign with everyone and we try to

53:39

experiment. So um currently I mean we

53:42

will soon move to discord. So we will

53:44

have a discord community for the

53:46

campaign. Uh we also already have a

53:49

desim page for some specific activities.

53:53

uh when I started coordinating the

53:55

campaign, I tried to co-create each any

53:57

feedback to the European uh commission's

53:59

public consultations together

54:01

experimenting with mentimeter or with

54:02

other methods for gathering collective

54:05

intelligence. Uh I mean we are trying

54:08

our way out because we do see that what

54:10

exists is not working and probably we

54:13

need to experiment and this is probably

54:15

what I like the most of this campaign.

54:16

We are literally building a political

54:18

pile from scratch and it's extremely

54:20

enthusi can say it's extremely exciting.

54:26

>> Absolutely. And thank you again

54:28

>> Thank you Alexandra for hosting us.

54:28

>> Thank you Alexandra for hosting us.