Daly: Hello and welcome to new voices in AI, the series where we celebrate the voices of students, early career researchers, and those with a new perspective on AI. I am Joe Daly engagement manager for AI hub, and this time I am talking to Nicolo’ Brandizzi about his work on collaborative human and AI systems. And without any further ado, let’s begin.

Daly: Hello and welcome. Thank you so much for joining us today and if you could just introduce yourself and who you are, where you are.

Brandizzi: Yes, thank you for having me. I’m Nicolo Brandizzi. I’m a PhD student at University of Sapienza, Rome, Italy, and I’m just in my second year of PhD right now.

Daly: Nice, and so how did you kind of get into AI?

Brandizzi: It it was quite unexpected actually, I just found myself in AI guess you know I was just following every courses and at the end of the semester usually have some motivational speaker coming from you know, master or also PhD and they talk about their life and I was pretty amazed the idea of doing a PhD and continuing my studies. Actually I was pretty scared to go work, so it’s like I love studying, let’s continue with that.

Daly: Yeah, that’s quite relatable. And so what kind of things are you working on at the moment?
I’m working especially on reinforcement learning in multi agent systems, my aim right now is on emergent communication, which is that kind of communication that emerges between artificial agents sharing a common goal in a game, you know it can be any game, it can simulate our real world environment or an Atari game the same. So and my actual scope is actually to have also human in the loop, you know a multi agent system with both artificial agents and humans and have a kind of common language to interact between each other.

Daly: That’s that’s quite cool, so it’s like kind of humans and AI kind of like interacting and like.

Brandizzi: It’s really hot right now, humans and AI.

Daly: Yeah, yeah it really is, and and I guess it is kind of working towards kind of like a shared goal, kind of in the game?

Brandizzi: Yeah, we have plenty of artificial intelligence which are way better than human at doing stuff you know. We nice and we lack the effort in order to have both human and artificial intelligence as agents to work together for a shared goal.

Daly: This sounds like a really interesting kind of area, but what are the kind of things that really excite you about AI kind of in general?

Brandizzi: In general, the one thing is, it really excites me is the ability for AI to show off the work of human beings. So we as human beings can be creative, we can shine in areas where artificial intelligence, at least right now, it’s not even comparable, you know. But there are also areas in which human abilities are kind of wasted for repetitive and automated work and jobs which can be automated, and I believe that having artificial agents together with humans can save both a lot of time and effort for a human being, but also can bring us more free time to, you know, focus on those kind of ability we have that agents, artificial agents, do not.

Daly: That makes no sense, yeah, and what do you think are some of the kind of big biggest challenges or sorry changes in in AI in the next kind of five to 10 years?

Brandizzi: Well, I just talked about, uh, job automation, which is a big, no big Red Cross for, you know like a lot of human beings, because they always think about robots taking over their jobs. But I mean the integration between both words would be the best for everyone. Also to you know, and I think this is one of the biggest challenge we have to actually embrace this kind of automation and also to understand that it can be for the best. Obviously we do also have to consider ethical issues and work related issues, but I mean in my opinion this would be the one of the biggest challenges in the upcoming years, especially considering our focus right now on human centered AI.

Daly: Absolutely yeah, it’s kind of the there’s like, I guess, so many aspects of kind of combining human and AI kind of abilities.

Brandizzi: Really hard

Daly: There’s a whole PhD on it.

Brandizzi: That it’s not going to be enough, not even close.

Daly: No (laugh) I mean, we kind of touched on it a little bit already, but what do you think are some of the implications of your research and sort of what kind of makes it interesting?

Brandizzi: Well, yes. As we already said, the implications are more free time for human beings. And I mean free time, for me, free time is actually activities or all those activities in which a human can shine you know. It can be a hobby, can be really good at sports, it can be really good at art and paintings. And also I would love to have a society where like that’s my personal goal (laugh) a society where you have a, you know psychologists and therapists being aided by artificial intelligence, where actually so right now you have one therapist that is following I don’t know 5 to 10 patients I guess that’s a that’s a pretty high number, but my hope is that with the aid of artificial agents you can actually follow more people in a more personalized matter, and also the important thing is here that you do not get rid of the therapist, that’s a must. You know, it’s a human being. The human beings cannot be replaced, but you also have the ability for him or her to follow more patients in a more personalized way, yeah.

Daly: Yeah, yeah, it’s really kind of supplementing rather than replacing is.

Brandizzi: Yeah, it’s not replacing it’s aiding.

Daly: Yeah, exactly and I guess you’ve also said already, just then kind of about the therapists, how they could potentially use it. But what are some, maybe other ways that you hope that your research could be used in the future?

Brandizzi: Other ways, well, I guess a therapist was an example, but you also have medical doctors. During the pandemic you had a lot of doctors kind of the using automation and actually robot the pepper robot, I don’t know if you’re familiar it’s like this really small and cute robot, which you can actually interact with patients being in intensive care hospital you know, being really sick of COVID or other diseases. And a doctor cannot be in the same hospital, but not even in the same country can be all the way around the world just doing its job, but again, aided by technology.

Daly: Yeah, yeah, I think the point about COVID is certainly very salient at the moment. How to use these technologies, especially in these kind of trying times.

Brandizzi: I zoom my best friend right now and I’m so used to zoom, Google meets and everything, on one side it’s really comfy to be in your room and in your house or in meets on the other side you do miss of you kind of miss the human interaction part a lot.

Daly: Yeah, yeah. You gotta, I guess you kind of you hope that there’s like a nice balance somewhere between the two, yeah?

Brandizzi: Balance is the most important in life you know, but it’s super challenging.

Daly: And so I I know, besides kind of just the PhD thing, everything that we’ve kind of covered so far, you are also involved in some other things. Can you possibly tell us a little bit more about some of those?

Brandizzi: Actually I’m involved in some organizations here in Italy, just like the Erasmus student network, which is super fun. I went in back in Erasmus in my master and it could never leave it, so I joined the ESN organization, which I really recommend for helping Erasmus students, but also this new research rising R2Net, which is a subset of Claire which I have over here, and it stands for Confederation of Laboratories for Artificial intelligence Research in Europe. It’s quite a mouthful.

Daly: Impressively done!

Brandizzi: And we do have this R2Net slack channel, where we, uh, I think we just reach 120 people and we are all PhD students from around Europe who are interested in artificial intelligence and researching kind of different fields. But we have a lot of meetings and also kind of fun activities, pop quiz and stuff so we can discuss more about our research and just know to and just get to know each other. We’re also scheduling an in person meeting, probably it’s going to be in few months if COVID goes alright so we’ll see.

Daly: Fingers crossed yeah, and we will also make sure to have some links to all of that as well on the website, so if people do want to find out more that’s easy, easy to do.
And and I guess to kind of well suddenly wrap up and so we have a chain of questions from the other interviewees in new voices, so we have your kind of surprise question from Maria de Arteaga who asked what would you like to see changing about the culture in the AI community?

Brandizzi: Hmmm, hard question about the culture in the community. I guess we like a more human centered, but we are having that already. Ethical issues are already being addressed, so I mean, ah, yes, actually, how do you call it? I would love to see more researchers involved in communicating science you know to people that come from other fields like or not even people that are do not know anything about artificial intelligence. My friends usually they think of me as the one implementing Skynet for Terminator and it’s quite upsetting sometimes because I’m here trying to do my best also to, in my opinion, help humans and humanity in general, so I would love to see more researchers involved in delegating their fields, their studies, and their research to everyone, yes in a more accessible way.

Daly: Yeah, and if I mean that’s actually kind of in the right place, kind of with the AI of one of our kind of main things is trying to improve kind of science communication because you say about Skynet and all of those kinds of things. I mean I’ve definitely heard that a lot of people saying kind of are you making a robot army?

Brandizzi: That’s the number one answer when I get to say.

Daly: Yeah, so science communication is I think it’s really important to kind of combat that as much as possible. And yeah, that’s a really great answer. And and yeah, really great answers across the board and thank you. Thank you so much for joining us today.

Brandizzi: Thank you for having me.

Daly: And yeah, that’s really great, thanks again and goodbye goodbye for now.

Brandizzi: What about the the question for the next one?

Daly: Of course, yes. I can’t believe I forgotten that. Yes, what is your question for the next person?

Brandizzi: I kind of have two, the first one is quite easy and it’s like, uh, would you rather go with academia or industry? So that’s one question everyone has in the PhD. But the other one, it’s like more if they answer academia, I would love to know their opinion on how it’s the research, you know the whole research thing doing you know with journals that PhD students must publish in journal and it’s a must, you know, and journals they get to be paid for that they’re not paying reviewers, so all the organization, what their what’s their opinion on the organization, not like an answer, and not expecting them to have a solution obviously, just an opinion.

Daly: Yeah, but I guess do is is it kind of like the the flip side of that question for like industry people?

Brandizzi: Yeah, distributed there, just it’s I mean. You know?

Daly: Yeah, that’s a really, really great question. I’m glad that we managed to remember to get that in. We may or may not have a blooper reel. This may all stay in.

Brandizzi: We will see we will see.

Daly: But either way, thank you again so much for your time today.

Daly: And finally, thank you for joining us for this interview. If you would like to know more about the series or find any of the previous episodes, they can be found on AIhub.org. Do join us for the next episode where I will be talking to Oumaima Hadjri, until then, goodbye for now.