Player is loading...
Panel discussion
Wednesday, June 15, 2022 · 5:30 p.m. · 46m 06s
Moderator : M. Nicholas C. Niggli, Secrétaire général adjoint du DEE, Mme Aurélie Rosemberg, Responsable de programme, Fondation Dalle Molle, Prof. Hervé Bourlard, Directeur de l’Institut de Recherche Idiap, Prof. ordinaire EPF, Mme Ioana Puscas, Chercheuse en IA, programme Sécurité et Technologie de UNIDIR, Me Nicolas Capt, Président du “Sommet genevois de l’Intelligence Artificielle » - AIGS, Dr Ricardo Chavarriaga, Directeur du bureau CLAIRE en Suisse, Membres de la Fondation Dalle Molle : Prof. G. Di Marzo et Prof. C. Pellegrini
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
when ladies and gentlemen thank you so much for for having me on because we agree on that
00:00:06
pretty general of the department of economic and employment even geneva
00:00:11
and i'm i'm going to facilitate the panel discussion this is often
00:00:16
um i'd like to start with with a few words uh uh
00:00:20
uh who which could be from a hundred low volume on it
00:00:27
the words are the following the quality or not the
00:00:31
longevity of one's life is what is important to set out
00:00:37
it's martin luther king nineteen sixty four up
00:00:42
his nobel prize winner and he was acutely aware
00:00:47
of the importance of a concept quality of life
00:00:51
and they're suffering dot coms from the lack of it
00:00:57
fast forward now to twenty twenty two
00:01:01
and you half exchange of bilateral exchanges between
00:01:06
prof like a nice which are going before going way and
00:01:09
still the focus is quality of life or the lack thereof
00:01:16
proper going east one comes up with a question
00:01:20
what kind of thing makes you feel fart or deprived
00:01:25
and the answer by protagonist to that's how i will all it is
00:01:33
yeah a lot of time feeling trapped and alone and having no means
00:01:38
of getting out of these circumstances makes one feel start depressed or angry
00:01:46
project on this one is a google and junior with now on gardening leaf
00:01:55
ah a proper gonna used to is essentially a google a high chop watch
00:02:02
and that just came out like three days ago without a sort of conversations that you have in between
00:02:08
a a a human intelligence and and they are you a good eventually and
00:02:15
this obviously poses quite a few questions really really really relative to quality of life
00:02:23
are we moving to words sentiment a are your is indeed a yeah
00:02:27
i no doubt smart i think we've we've worked out this afternoon uh um
00:02:33
what would it mean if a i would become french and um and
00:02:40
and then perhaps what can we do to make sure technology and especially a
00:02:47
you are in service humanity rub it in your project and to complement this
00:02:54
when it comes to pushing the boundaries of innovation uh for humanity should we not rethink about
00:03:01
the importance of a a are you and teach them comments in the spirit
00:03:06
of the work uh don't by professor ignore or from the two thousand nine
00:03:12
no a nobel prize winner for economics the first one we got this price and i
00:03:17
think it's it's more about the nobel prize and the quality of women in in economic research
00:03:24
so these are some of the big questions which i will explore with a collection of brilliant minds
00:03:29
and heart we i think we need both a present
00:03:33
here on the panel today and it gives me now uh
00:03:37
a great pleasure to to call on on the podium uh a few people but i think i really have to here
00:03:44
so so i ladies first overview wasn't about a
00:03:49
superwoman a responsible for the fiftieth birthday of dudley moore
00:03:54
a project manager at the foundation thank you so much already for being here
00:03:59
then i have an equal our cart founding partner of captain these i
00:04:03
think that's that's correct that's the name of the low for men you're the
00:04:06
president of that was the previous one but the only reason was good memory
00:04:10
yeah and also the okay you will tell us about about the new one
00:04:14
uh and president of the geneva also made on a high and i'd like to to our to this this
00:04:19
collection of great minds uh every all our where is
00:04:23
every uh every uh come to the stage a very uh
00:04:28
every you're you're the director of the are we we spoke about you're really in the situation right before
00:04:34
professor at the pure further as well and board member of the valley
00:04:39
more that foundation then there is he still basically anywhere is his you'll
00:04:44
here we go because your is your your your your of your former
00:04:48
director of of of uh the sort when you have the or format peak
00:04:52
and the board member of the valley more information as well then there is
00:04:55
your now who's got apologies if i use sort of miss we sounded your name
00:05:02
but i i think we got it your does your senior a guy restructure the
00:05:05
united nations usage for disarmament thank you so much for being with us uh you wanna
00:05:12
rico chablis the ah actually at our uh uh
00:05:16
we cover also house many heart uh he is the
00:05:19
director of history spiro the configuration of rubber trees on a i research in europe but you're also work
00:05:25
probably marked for what pages use p. uh you you might even to tell us what this means
00:05:31
i don't think you really go for the year and then there is
00:05:34
juror no uh uh the market so professor and director of this you wild
00:05:39
and also board member of that probably more the foundation so i'm gonna i'm just gonna you get
00:05:44
away from here and and give you give you my spock please or call come come closer so
00:05:50
i'll give you the floor intern i think that's the
00:05:53
the biggest panel i ever tried to manage uh almost impossible
00:05:59
uh so but before giving you the flow we have about thirty five forty minutes
00:06:03
together and and so i have to tell you the the rule is pretty simple
00:06:08
it's p. c. p. when she comparing and to the point please
00:06:14
uh because we all have to time to uh do anything else and
00:06:17
often i'd like to start with with with every uh every uh well
00:06:22
to the point yes absolutely how to see the evolution of a yeah research in the next
00:06:28
few years and what you think other fields
00:06:31
where a i will make the greatest advances uh
00:06:38
uh
00:06:40
yes many families of the ah there is really hard to to answer the question so just
00:06:46
what is certainly there's also this model only i know as opposed to the old i guess
00:06:51
what is the main difference between hold a uh yeah called expert systems in the fifties
00:06:56
and on into the new idea model yeah there's not much you know besides digital data
00:07:03
all the data is digital what for many of us in industry when you would win just it is
00:07:07
like eighty percent of industries where they still don't know what digital data is we can come back to that
00:07:14
something you know so many families of yeah but i would i would classify
00:07:18
them into to to simplify the discussion to be funded the photos one is
00:07:24
trying to do things that human beings are not able to do at all
00:07:29
processing huge amount of data to extract
00:07:32
properties and help humanity quality of life
00:07:37
and so and so was a huge amount of data the second
00:07:41
family i'm simply over simplifying the whole thing but the second families
00:07:45
yeah i tried to me make human beings skills
00:07:49
or augment this is a funny development human beings could
00:07:53
in the first case i think this is with the big potential is into medium term i would say
00:07:58
a high processing a huge amount of data i'm a little bit conservative but they also
00:08:03
like to be a little bit provocative in one place it before people will do me
00:08:09
i believe that if we gave all of us
00:08:12
humanity give to write to explode all all data
00:08:16
that we are generating every the massive amount of data that we all of us we are generating everyday
00:08:22
if all the humanity would agree to she a hundred percent one hundred percent
00:08:26
of the data they usually they i see all the legal people you know
00:08:30
starting shouting it means i'm sure we all their data i'm show that humanity would make
00:08:37
use progress we would understand social behaviour or the the the climate change will
00:08:44
will ecology we would be able to what you do data mining in that big
00:08:50
so we are going much beyond what human will ever be able to do and we would make
00:08:54
big big progress in quality of life you may need to install i'm convinced of that and this is within reach
00:09:02
now people are going to a hundred percent of the uh yeah
00:09:05
i agree on my site i am i start being fed the
00:09:09
bustle of each of us we are a bit further i know
00:09:13
like taking this little box each time now we access any stupid website
00:09:17
although this confidentiality box you know and private issue that we understand
00:09:22
all the things and so on we checked the object of boxes
00:09:25
right all the time we're all funded by that because they want to access that for a new website
00:09:32
only side i would like to have a system where i can describe my profile
00:09:37
what am i had a ready to share what and what we did to share yeah and so
00:09:42
and then people will know what website would known from that automatically first point
00:09:48
then i would like to be we've paid for what i'm sharing
00:09:53
i don't believe i don't understand and that's something we should regulate that only a few
00:09:58
companies making money based on top of the data did we generate we are the data generated
00:10:05
we out of the owners of the data and only a few companies to the ah making profit
00:10:11
from the data i generated i want humanity to be
00:10:15
able to make money i want to date other to own
00:10:20
and based on the that the closer to the to the level of shaving you know
00:10:25
between zero percent and one hundred percent i would like to be able to get money
00:10:30
four percent the function hundred percent of my data i would get something like
00:10:35
which is called is the is the universal sorry why not two thousand francs amounts
00:10:42
if i should only some really thought you any fashion zero percent i would get
00:10:47
no money but it's okay right i think there is something that we can really today
00:10:52
okay then in the case of the older yeah i yeah we're we're using data to recognise images
00:10:59
into to do speech recognition to do translations on i don't think we should expect much more problems
00:11:06
except if we fall cues oh yeah i used
00:11:09
to expand human skills not to replace you you know
00:11:15
every i gave told you know i i asked people to raise yeah the hand
00:11:18
if they still believe to suit on the most cost before the end of the year
00:11:22
i've been doing that for ten years every uh uh the
00:11:25
fewer and fewer people within the last conference id force in the
00:11:29
in the rose then yeah i'd want a buck meeting board meeting
00:11:35
only one person out of two hundred posted wheezing today and one and nasty and that would be next dealing with zero
00:11:43
so i don't believe in magic i think it's just hard work digital they tried to call this the main
00:11:49
difference between the old yeah and the more than yeah but this one that we can can can go for it
00:11:55
thank you so much uh uh with a messy because it's uh i don and
00:12:02
and prices power but i caught on and then obviously sort of reversing this uh
00:12:08
detective digital feudalism kind of uh uh you know triangle power to to to to be
00:12:15
on call that i was you i am i'm very conservative but i am hundred personally either
00:12:21
okay yeah i i like to work transparency item we need new full transparency
00:12:26
on both sides talking about transparency of the the topics probably just to keep the company's
00:12:33
explode into data i let me i think we need that's what transparency from the data providers
00:12:38
i want you know i think we need to be transparent off almost data we're willing to say
00:12:43
uh_huh and if you're willing to show that they don't have to be paid for their end users of the data
00:12:48
they also have to be transparent about what they do with all the love and the yeah i
00:12:52
can also help us to build tools that will follow that will track to you like this stuff
00:12:58
there is a lot of progress being made in by the matrix yeah is insecurity i think we
00:13:02
can use exactly the same technology to track what people is uh people are doing with my data
00:13:08
people talk a lot about the the biometric possible it's very easy to get a very estimate as more s. m. s. message
00:13:15
each time someone access you by omitting possible and you know it's like looking into a by
00:13:21
the window or something right that if you know that people we know that you are looking
00:13:27
by a little closer to you and you know if you don't do it anymore
00:13:33
i got a thanks terrible lot thank you for the last uh uh uh historically that move to you
00:13:42
draw points often used to save up money is nothing else but what he
00:13:46
makes of himself now humanity has also
00:13:50
started to create artificial intelligence not just
00:13:53
human intelligence how will artificial intelligence contribute to the quality of life in the next fifty years
00:14:01
again we you know fifty fifty hours a i in three minutes obviously
00:14:08
yeah oh yeah but mulder
00:14:12
thank you very much
00:14:15
i'm going to answer this for difficult question in french to uh
00:14:20
give everyone a break it so how can artificial intelligence
00:14:24
contribute to the quality of life in the next fifty years
00:14:30
well obviously uh we don't have a crystal ball uh tell disposal
00:14:34
so it's going to be very difficult for me to answer this question precisely to move the less
00:14:41
after having discussed
00:14:43
it with some colleagues they told me refer to your personal experience so with your permission
00:14:51
as of overdue lifter for several decades and been doing research in
00:14:55
this field for several decades i remember that when i was a asked
00:15:00
this question at the beginning of my career what do you do and what subject you work on
00:15:06
and if a a
00:15:09
replied referred to upload differ replied to artificial intelligence
00:15:14
the interviewer would to the give me a funny look and say uh how sales think how can one
00:15:21
work on such a strange subject uh which is rather rather obscure and there's not a very use to anyone
00:15:29
that's true that since then things have evolved a lot remember um interesting episode
00:15:39
when with professor for things of yours and the purpose of a
00:15:42
fulcrum zurich when we were submitting some projects to the swiss national fund
00:15:47
it's as soon as if we received some rather skeptically replies we had
00:15:52
to go to bed and several times to try and convince all all the
00:15:59
uh partners that it was not to strangers that and that this kind of thing was done
00:16:02
elsewhere also it's what's in them there was no reason not to do this since woods and
00:16:07
now if we look at the situation we're in right now things have changed a lot fortunately
00:16:15
i is everywhere you have it in your dishwasher you will soon
00:16:19
have it in your car to drive you where you want to go
00:16:22
automatically i will have it and this is being referred to but this because
00:16:28
when you go through the new automated possible chick at geneva
00:16:32
report they've introduced to new device you place your devise a
00:16:37
on the the plate to um the the system uh if you take such a possible lets
00:16:44
you through what's going to happen in the next fifty as i don't know on the other hand
00:16:51
if i turn to the person whose portrait it's on this wall angelo to label it
00:16:58
and if i consider
00:17:02
what his motive and the result the troll was
00:17:05
a two launches foundation like to go to if uh
00:17:10
the advances of technology i. t. overall politics so and uh i uh to
00:17:16
serve human beings and if human beings are not to be in sleep like technology
00:17:21
if you much of the sleeper technology then things will be heading in the right direction unquote
00:17:27
now how can things change and continued to change in the right direction in the next few
00:17:33
years prior to that i will turn to the second poster which you can see on the wall
00:17:39
and i'm convinced that the quality of life for label which is going to be launch
00:17:44
but the foundation will play an important role why because it places the emphasis on the
00:17:52
number one am of missed the telly molly and at the same time this label will
00:17:57
uh trusts a broad range of people researchers companies industrialists
00:18:04
all these people are going to be people if they so wish and in
00:18:08
compliance with listed a little is we should be able to positively contribute to
00:18:16
artificial intelligence being at the service of human beings
00:18:19
and not vice versa thank you very much indeed to
00:18:25
so there's a from mister medium thank you for breaking down the wall between english and french
00:18:32
speaking french uh so um i will try to avoid speaking a swiss german the stuff to do
00:18:41
yeah
00:18:43
let's get back to to english you're
00:18:45
probably christian mention cars and the intelligence in
00:18:50
vehicles and obvious that the transition i'd like to to to to take
00:18:56
when the system is made of a when a system is made or tournaments thanks to a yeah i
00:19:04
we generally find an excessive trust of the human two words information transmitted by the
00:19:11
machine and therefore potentially a loss of concentration
00:19:16
of cognitive attention uh by the human being
00:19:20
how do you see the the future about human machine interface
00:19:25
uh uh and and can we eventually do something to maintain the human cognitive involvement throughout the
00:19:31
use of the system you know terms uh how to make a yeah you uh uh uh
00:19:37
sort of getting people to to to get easy if i could sort of simplify my own question please you all know
00:19:51
i think you you you voices to be heard of something i think is the different question actually and
00:19:57
i think that we're finally getting a bit closer with this question and so on
00:20:01
making a ice nadine or e. s. l. let's think a bit about what it means
00:20:06
really for humans interact an autonomous system because we talk about yeah in software about um
00:20:12
i think that processing about things it can do in in the
00:20:15
end we consider computer but what does it mean we interact but uh
00:20:19
with the system that is upon an answer it has a high level of of automation
00:20:23
and and the passion points to loss of
00:20:27
attention to bias that that humans tend to have
00:20:30
when they they live for for a long time on a system that works well that is
00:20:35
i've had a certain level of predictability and that's a real challenge but i'm just going to
00:20:40
to start a bit with the background before answering that even though you said to be at the uh to the point that i
00:20:45
would be to the point and i think the first um if i shall engine here there's more and more studies in the is on
00:20:53
um human machine interaction in the context of autonomy of course autonomous card obvious don't exist
00:20:59
not not at the highest level i think that the stakes of autonomy that doesn't exist but
00:21:04
there's the high level of about autonomy in certain systems in in cars even is that certain happens is then
00:21:10
so in that situation when we deal with says they have a high level of autonomy
00:21:15
it's been shown time and time again that after a while the seven nine twelve seconds usually you can
00:21:21
i started his attentions all these that relied on this is the nannies especially if it works very well
00:21:27
for a long period of time more and more that attention is lost and the waitress that we tried this is then to
00:21:33
navigate while to take as the places or to just decide to
00:21:36
to make the right decisions because we just like to trust it
00:21:40
not the problem that was it this problem and to sort of come
00:21:44
it is the challenge was up again through um a sufficient into interfaces because
00:21:50
the interface is no way the meeting point between the human and uh
00:21:53
this is then sell looking at designing interfaces that there are making them more
00:21:59
not just user friendly but more uh designed in a way that
00:22:02
would cap which is the key the human involved engage in there and
00:22:06
the process of using that since then has been fun as a sort of a solution but then even they are actually
00:22:12
there is very little that we can do just to design itself it with the design of an oven interface because
00:22:18
human attention i think the car that will have more to say because i don't get the neural scientific background yet
00:22:23
that it's very it's very hard to to maintain for a long time no
00:22:26
matter how well designed and interface itself there's this this is a fundamental problem in
00:22:31
human machine interaction in the context of autonomy that no matter how how well we design an
00:22:36
interface loss of attention and and that sort of a can complacency which is kicking at some point
00:22:43
and then there's that other solutions through for example interstate interfaces where
00:22:47
well you i haven't more realistic experience of that of the use of the system of the environment but even
00:22:53
there of course that there there is no magic there's no silver bullet for and for keeping semen actively involved
00:22:59
while still it is really the challenge of course in uh in in
00:23:03
using cars but even more something but because i think you and um
00:23:07
now working hours that is autonomous weapon systems but the consequences of course i
00:23:11
am much more stevie and when when dawson attention cars and i says then
00:23:16
arsenal function and then it's very difficult for human to run up their alertness in a moment of prices
00:23:23
so that the answer to come back you know in a circular ways that in a way we are
00:23:27
yeah really facing the challenge in in human machine interaction in this regard because it's very it's still
00:23:32
very hard to calibrate attention an expectation interest when
00:23:36
we use when we get to take it on assistance
00:23:40
you as a forward just a lot do do we need to develop use keys in humans i
00:23:45
mean we we we speak a lot about this actually the we we humans need to our future work
00:23:51
do we need to to learn new skills uh when it comes to interacting with the i. e. what your take on that
00:23:57
oh absolutely we need to learn with that an excuse and also
00:24:00
new training nap it's when when mean gauge with an assistant self
00:24:04
for example this is done and i read recently study on um that they slammed um it does
00:24:10
not driving experience where you have to update uh of course there is it's provided there is um
00:24:16
there is an updating the software constantly because the system keeps learning and then you
00:24:20
how you have that data that uh that's training you as it is the driver
00:24:24
the saints things are now beginning to i starting to be studied in that
00:24:27
field of autonomous weapons systems were looking more and more at what training can
00:24:32
can but that's it entailed forces stands that keep learning and we
00:24:36
have to keep updating a mental model of the system you're using because
00:24:40
this is then you start using today might be different in a way it learns or the weight behaves in your six
00:24:45
months from now sell training has to change it would and that requires obviously much more agility and with the war that
00:24:52
everybody uses now agility agility but it's really it really requires a lot of sort of uh
00:24:58
um and preparedness to to learn constantly and to update that that training more regularly
00:25:03
uh_huh thank you so much you got on the you heard my my starting point
00:25:11
these really are essentially not being very happy about its its condition
00:25:16
do you see yourself defending a an artificial intelligence one they don't sort of the first part
00:25:21
of my my question and and second really how how do you see a uh this this
00:25:27
this agenda of of digital commons and implementation of it uh on
00:25:32
a on a global level i mean you think it's gonna calls
00:25:35
legal problems and what can switch didn't do it on top from thank you so much
00:25:41
and as apprentice only see the point of course i personally difficult for like the so i'm not trying to answer in french
00:25:48
so first first question i think that to defend yea i what is interesting
00:25:53
is that today we haven't tackled that then interesting question is that i i
00:25:58
a i is a intelligent or not because this is like i mean it's in society that to an average a ooh
00:26:05
ooh citizens say uh the uh a i would replace a
00:26:11
kane and in fact it's just there to improve
00:26:15
with data set up to improve the augmented reality so
00:26:19
we can see that uh we are back to the ground and in the end we don't speak a lot about uh
00:26:25
the uh strong guy uh artificial intelligence like a fantasy which
00:26:31
ago was the further and further them all go forward stop
00:26:35
this uh europe and uh the draft opening act is quite pragmatic that was a conference is
00:26:40
speaking about uh the the regulation of uh what we cannot predict and this is the case so
00:26:46
to me this is intelligent this is clever the way it was made uh they are summaries cousins and
00:26:53
a list of what i wanna be happy with that
00:26:56
but uh a biology in public spaces interesting in that respect
00:27:01
and uh if it at some point we put all our
00:27:04
data as somewhere and this you question i think because that
00:27:08
we always say that uh d. r. g. d. g. p. r. is very interesting et cetera
00:27:13
but uh we every time we need to put a disease across
00:27:17
but philosophically it goes beyond its means that uh we want to
00:27:22
to put today test subject uh in the focus somewhat these data but we never
00:27:28
it it it it was it was never it in in the in the silly that we don't
00:27:35
want to beat i mean i thought because well we always asked to be our profit also anyways
00:27:41
we'd all that's so that's so nice about if it's a
00:27:44
she like that which don't exist so why now would die defence
00:27:49
would i defend a a i i'm not sure i did that in this yet so it was quite funny but i mean
00:27:55
these being said there's some quite outstanding things uh there is
00:27:59
the uh picture a movie made by job the initial but well
00:28:03
where it was quite good so it might occur but i'm not then a a
00:28:08
next uh five years autonomously hike or they're not here not now it's not very clear
00:28:15
and so uh yesterday i went through the mit
00:28:18
more machine i mean we have to me in and
00:28:22
i don't know that's exactly how i can we have to decide whether we want a um is a
00:28:29
to have a crash show with that dog oh is that with uh something gets and then in between
00:28:34
a bikini or are docked are or something like that and the choices i mean that are just all four
00:28:39
i mean they just so i go for a mix of everything
00:28:42
and we cannot do that i mean it is very interesting and it's
00:28:49
but even though we won't have main maybe
00:28:53
autonomously hikers or smart machines that we'll be some
00:28:56
difficulties in their regulating and as for the digital colours i don't have much to say in fact
00:29:04
as soon as we want to put in place something and globally yeah and we need international
00:29:09
conventions which are quite a baby to implement and
00:29:14
dead don't really cover a lot of things so
00:29:18
of course to me the outcome is quite good but if we go beyond
00:29:23
the united states it's they are quite weak tools in the end of the day
00:29:30
thank you
00:29:32
i think in colour
00:29:34
the capital mm should be uh uh
00:29:39
continue on this very front of is a yeah you
00:29:43
small or not and i heard nicholas and i are very
00:29:48
if one looks a history uh artificial intelligence house probably so for bin
00:29:56
best interest and not true stupidity
00:30:01
uh do you agree with this statement first and how
00:30:05
you see the evolution of the modulation around a are you
00:30:09
and a given your background i'd like to hear you are
00:30:14
saying a few words about the necessity to be more trans disciplinary
00:30:19
in the approaches we're taking on on on research but also in
00:30:22
terms of implementation thank you so much legal thank you very much
00:30:28
so point she sighs it's i think there is nothing
00:30:33
more dangerous than human intelligence or stupid even g. day
00:30:38
all the race that come with that vision takings are
00:30:41
human eight hours human generate so the rest started human
00:30:46
and we should probably avoid the temptation of talking about and a i
00:30:50
as an entity in itself there's multiple instances of a.
00:30:55
i. systems or systems that exceed eight a. i. characteristics that
00:31:01
still not an unique and it yes and a high maybe something that we have
00:31:05
to keep in mind when we think about how do we establish governance for a high
00:31:11
because there is not a single entity there with people ways we're
00:31:14
implementing it with people ways for developing it would be ways for
00:31:18
using it and what people ways for our putting an oversight system
00:31:23
thought if first point we need to to address and there are
00:31:28
again years risks human made stars human made solution should address a human
00:31:34
aspects of developing the i deploy in a high and reacting to usage yeah
00:31:40
if you look at how are we developing a legislation
00:31:45
will expand it to governance and the path we should follow
00:31:49
i i would say that uh it's an impossible task we're trying to do
00:31:54
something here that cannot be she to call we if any
00:31:58
fixed ruled that will solve all the problems i would let
00:32:02
us be happy uh let's see how i stop worrying about
00:32:05
the phone but god that's not possible but it's not a
00:32:10
task like us is divorced one where i just push the rock all on top of
00:32:16
a hill just put it down because says it was going something that had no point
00:32:21
we're doing something it is very hard but we can achieve because there is no point to it
00:32:25
but it is worth doing and is gonna be
00:32:28
and never ending process of starvation rules and governance mechanisms
00:32:33
that are necessary to reduce there is that these technologies and tail these technologies that are
00:32:39
emerging that are normal that active will think so oh we come
00:32:43
back to this analogy or trying to fly thing why rabbit and
00:32:47
when we're doing that we know that we don't have all the
00:32:50
necessary information to set all the rules to set the strict rules
00:32:55
ah but we cannot wait too long to gather this information in order to set
00:33:00
before so we need to play with these different a time frame is we have to
00:33:04
also take advantage of the increasing it is the knowledge that we keen as we see the system is being used
00:33:11
and that is the thing to something that was referred before ah we she's the different colours mechanisms and
00:33:18
tools we should not focus only on regulation only on
00:33:23
particularly got use remains because they yeah they full feel
00:33:27
one role among the different techniques we it was mentioned that if we want to half
00:33:33
a lot well treaty a global framework then we
00:33:37
need to reach consensus with many many stakeholders that
00:33:44
usually takes for ever that ends up what's going down some of
00:33:49
the more air choose your parts and mortimer strongest part of this
00:33:54
saw that isn't this model is not suitable for setting the rules for
00:33:59
everything related yeah it's a lot of value of doing that because that's that's
00:34:03
they generally ground but then we have other mechanisms need to
00:34:07
be in place that need to be coherent with these general consensus
00:34:11
that allows us and empower us to do and apply these uh
00:34:16
these mechanisms and here's where we're talking about holidays binding last also
00:34:20
just awful mechanisms to stand artsy look at the the way they
00:34:23
act it's relying strong on standards standards that don't exist yet so
00:34:29
this is something that we need to to look at and don't lose sight on these
00:34:32
different pen is that we need to develop and we need when we talk about me
00:34:37
he is working to disciplinary works ah often we have in in events like that like a a
00:34:43
one session that goes on the technique one session that goes on daily aspects one
00:34:48
session of both in the attic aspect and we're still working on files and we
00:34:53
need to start increasing literacy all across the board with only dress in a i
00:34:59
guess a lot of energy on saying we need to teach people how to program
00:35:04
there is a lot less saying we need to teach engineer is how
00:35:07
to develop a legal framework or a common frame and is something that you
00:35:11
because if we we we want these engineers the developers to
00:35:16
take that these elements into account if we want an organisation to make decisions about
00:35:20
informed they need to have these different endings and this is something that is very important
00:35:25
two aspects that i think are necessary to reach a proper norman something like the first one is
00:35:30
empowerment like developing these capacities increasing uterus at the
00:35:34
second one he's incentives you want to get by
00:35:39
if you need to generate incentives for people to do things
00:35:42
in the right way to develop a pair that meeting responsible way
00:35:47
every frame or that we set in terms of legislation in terms
00:35:50
of the standard in terms of a binding well or so awful
00:35:54
needs to be outline somehow always be compatible we if the needs with the
00:36:01
rules and we think norms that a a certain social economical uh environment pass
00:36:08
otherwise you may end up with things like that that are not applicable or what
00:36:11
we would be impossible to generated by necessary to adhere to discuss these uh these technical
00:36:18
thank you very much indeed
00:36:20
incentives and empowerment that's where we go really uh
00:36:25
uh this more or less what we used to uh
00:36:28
capo also supports a a lot they're like they're not to yourself if
00:36:35
can you tell us something about this uh project to develop a label full artificial intelligence
00:36:42
that seems to be an extraordinary project uh can you tell us something about the time line
00:36:48
uh i think everyone in the audience is looking forward to seeing that label ready and
00:36:52
to put in place or what are the obstacles that the foundation will have to overcome
00:36:59
i'm in cooperation with the various sins compression with diverse but this one really
00:37:07
miss if you would introduction this is a label which we're launching
00:37:10
with the uh the foundation but also together with s. m. u. use
00:37:18
so it's often difficult person is to integrate to artificial intelligence in the cool business and uh
00:37:25
we come into quite a lot of we run up against a lot of restraining factors uh when we try to
00:37:31
do this with financing and the understanding of the ecosystem
00:37:35
rules that we have to comply with as regards this the
00:37:40
um approach that the foundation is adopted to create things
00:37:44
label isn't intuitive one the idea is to create to uh
00:37:48
uh technical tool which will comprising contain like a a a
00:37:53
a or a or a i rather and uh which will
00:37:58
control the way in which areas use by the eye enterprises um
00:38:03
which will involve ethical standards so as you can imagine it's pretty challenging
00:38:09
but i believe um our uh research is a tele million enough to be able to take up this challenge
00:38:20
assume is need to innovate to stay in the market
00:38:24
term it's difficult for them to stick to step because um
00:38:30
many of the things they have to learn of a very difficult to learn so first
00:38:34
of all we have to really challenge the channel this a research project in the right direction
00:38:41
we have a reflection group which consists of uh some very big
00:38:46
names in research and some free time to people so with first
00:38:49
will reflecting on how we're going to put this project in place
00:38:52
then we will try to find some companies which sure agree to
00:38:58
work with c. i. e. um to d. up the first stage
00:39:05
it's difficult to give you a timeline uh
00:39:08
i don't want to create false expectations but um
00:39:14
and this is something which is emerging from the ground on
00:39:18
the occasion of this fiftieth anniversary of the tell him only foundation
00:39:23
and i'm sure that the next few years will come up with something concrete to disregard
00:39:29
super but so is the moderator
00:39:33
well them with a a show from uh who was
00:39:37
already a set up the potential of c. c. you buy
00:39:44
and then a great memory earlier what is the potential of a c. u. i. to implement this technology
00:39:53
can you hear me
00:39:56
so um we were asking about the challenges and the potential yes
00:40:01
says that you're from that well we aren't into faculty centred universe
00:40:05
to for to me for this is both a challenge and the strings
00:40:11
so the structures that were of faculty but uh we use that we have
00:40:14
is that we are a centre which is supported by the wreck to rate
00:40:19
but at the same time the universe to seem to look up human
00:40:23
governments is such that often we are limited in what we'd like to do
00:40:28
and of course we are accountable you this too so the budget is what it is we cannot
00:40:33
develop it anymore well we uh we can try
00:40:37
to developers additional interesting projects like the one we are
00:40:42
running with to tell the deli molly foundation now as regards the challenges at the moment
00:40:50
there is a growing awareness at the level of the can turn on the region
00:40:54
that something has to be done in the area in the digital area and to
00:41:01
for a enterprises so there's a deer of uh developing research and
00:41:07
teaching in this area and to develop for digital technology even more strongly
00:41:15
so this is what i would like to see happen i would
00:41:19
like to see really strong signal being given by our authorities for example
00:41:25
to develop a digital technology most strongly in the can to
00:41:29
move to move out of this this is part of your objective
00:41:33
e. edge of time like the present extensively that don't you yeah i think you can
00:41:45
how to keep its uh switches to
00:41:52
stating in this record to want to the next steps um
00:41:59
what to the next actions that a counter geneva is going to be doing in the field of uh
00:42:06
artificial intelligence also thank you for the challenger
00:42:11
the i would like to give you a round of applause full having a state with the some tool half the time
00:42:18
for us a high is part of the broader feel does out minister said earlier this afternoon
00:42:27
namely the field of of placing digital technology that the service of all the
00:42:32
strong him forces of achiever and also mall globally that i'm
00:42:37
the regional level and at the level of the international geneva
00:42:42
i'll action uh can be situated at the transfers soul uh
00:42:48
a lot of these between these two axes as regards economy and employment
00:42:54
the reason beefy a witches materialist
00:43:00
with the consequences of the covert nineteen pandemic with
00:43:04
regard to the digital divide this is something which um
00:43:08
does still exist to very much in companies in a semi is you mentioned earlier the problem of following
00:43:15
all the regulations the problem of knowing what one is allowed to do what one is not allowed to do
00:43:22
especially for a a semi use which uh make up a considerable portion of our economic fabric
00:43:28
so this is a major problem them if i add to that the problem of cyber security
00:43:35
we'll see the uh mister a little soft and spoke to about a digital uh
00:43:41
confidence and trust and a i of course uh is integrated in that
00:43:46
we have the local regional agenda which is pretty strong in this regard as a number of initiatives which are
00:43:54
being run by the department to increase our footprint in this regard
00:44:00
what makes geneva really original and i devote to quite a lot of time
00:44:04
working on this is to um build the bridges
00:44:08
uh between the terms which ricardo refer to moments ago
00:44:12
trance discipline or t. v. or breach between a
00:44:16
digital components um small local forces speeded to um
00:44:21
the university of the uh couldn't hear him here we're in the process of the c. u. i.
00:44:30
and the h. e. g. is uh just nearby and h. e. s. s. o. is nearby that's
00:44:35
the university of applied sciences western uh switzerland and
00:44:40
uh the e. p. f. l. this was instituted
00:44:42
swiss federal institute of technology lows and also has um a brunch in geneva so um
00:44:50
we are very fertile ground in economic terms so
00:44:53
i'll challenges to get all the surfaces to work together
00:44:57
and to think and rethink these documents related issues so
00:45:01
we positional so so different to between these two uh
00:45:05
uh thrusts the local aspect so's not
00:45:09
to leave anyone behind resume in the ditch
00:45:14
oh and then there's a whole um acceleration that is required to to
00:45:19
a home school so we will need for future and to establish um
00:45:25
actually for a as a real the hub for a digital skills and
00:45:30
competencies it's about all i can tell you in the space of two minutes
00:45:35
but before
00:45:37
asking you to give a round of applause to the other panelists we
00:45:42
palace would like to applaud the panelists who have stayed home until after time
00:45:48
well the audience can have close the panelists intern
00:45:52
well i think we have a slightly exceeded the time it was a lot to to this panel
00:45:57
really the president love very dearly molly
Conference Program
Welcome words
Aurélie Rosemberg, Prof. Antoine Geissbuhler, Vice-recteur de l’Université de Genève
June 15, 2022 · 4 p.m.
Conférence opening
Mme Fabienne Fischer, Conseillère d’Etat du canton de Genève – Département de l’Economie et de l’Emploi (DEE)
June 15, 2022 · 4:05 p.m.
Evolution of Computing and Vision
Prof. Giovanna Di Marzo Serugendo, Directrice du Centre Universitaire Informatique de l’Université de Genève et Professeure ordinaire à l’Université de Genève
June 15, 2022 · 4:15 p.m.
Reality, Truth et Artificial Intelligence
Prof. François Fleuret, Professeur Ordinaire, Directeur du groupe de recherche en apprentissage artificiel, Université de Genève. Professeur Titulaire, École Polytechnique Fédérale de Lausanne. Chercheur externe, Institut de Recherche Idiap. Cofondateur Neural Concept SA
June 15, 2022 · 4:30 p.m.
120 views
