INKOVEMA Podcast „Well through time“
#202 GddZ – Digitised empathy and mediation work?
New technologies for mediation and conflict counselling. Part 7
In conversation with Michael Lardy
Michael Lardy: Mediator from Salzburg, Austria. He completed a law degree at the University of Salzburg from 2020 to 2024. In addition, he completed a diploma course to become a mediator at Wifi Salzburg between 2022 and 2023. Lardy specialises in family mediation and divorce mediation. He has also worked on the topic of artificial intelligence in mediation and has published an article entitled „Mediation and AI: The Silent Revolution – How Human is ChatGPT?“.
Well through time.
The podcast about mediation, conflict coaching and organisational consulting.
Contents
Chapter
0:03 – Introduction to the new technology
0:57 – Exciting developments in mediation
1:50 – Michael Lardy on the podcast
2:28 – The influence of AI on mediation
6:55 – Empathy and AI: A critical view
8:41 – Humanisation of technology
11:16 – The role of language in communication
15:10 – Trust in AI: opportunities and risks
18:40 – The future of AI in mediation
21:21 – Vision of an AI co-mediator
24:58 – Open LLMs: The next revolution?
28:22 – Requirements for future AI models
32:14 – Low-threshold access to mediation
36:13 – International perspectives on mediation
39:50 – Conclusion and outlook for the future
Summary of content
This episode of the podcast "Gut durch die Zeit" is all about the integration of new technologies in the field of mediation and conflict resolution. I, Sascha Weigel, and my colleague Dr Frank Termer reflect on the challenges and opportunities associated with the use of Artificial Intelligence (AI), in particular ChatGPT, in mediation. We discuss not only what is technologically possible, but also how such innovations could change the realities of mediation in practice.
Frank and I open the conversation by discussing current events that are influencing not only the political landscape but also our work as mediators. We emphasise the need to respond proactively to upcoming challenges without losing sight of the rapid pace of technological development. Our guest is mediator Michael Lardy from Salzburg, who gained early experience with AI technologies. Michael brings interesting perspectives on how artificial intelligence could revolutionise the way we resolve disputes.
A particular focus is on the question, How empathetic AI is and whether it can come close to human empathy. From a scientific perspective, we examine various studies that show that AI is capable of acting empathetically, at least in the perception of users. This leads to the important question: Are we humanising the technology and projecting human characteristics onto it, or does AI actually have the potential to act empathetically? This discrepancy is a central point of our discussion.
Michael Lardy shares his experiences from his practice and emphasises the potential of using AI in mediation, such as the possibility of going through role-specific training through AI-supported simulations. We also look at how AIs could facilitate access to mediation by acting as a first point of contact and helping people to better understand and navigate conflicts in advance.
In addition, we discuss the challenges and concerns regarding the ethical use of AI and the need to create the right framework for AI to be accepted in mediation. We analyse the extent to which mediators should retain control over the technology to ensure that the human connection remains in the process.
Finally, we discuss how important it is not only to be consumers of technology in the context of digital transformation, but also to actively develop strategies to optimise our own skills and those of the technologies. We educate ourselves together in order to better shape the future of mediation and suggest that low-threshold access to mediation should be opened up through AI-supported solutions.
This episode is a fascinating dive into the development and challenges that AI brings to our industry. It is thought-provoking about the role of technology in our work and how we can use it to benefit those seeking support in conflict situations.
Full transcript of the episode
[0:03]
Introduction to the new technology
[0:00]Two Aspects. The one, what is in the Mediation real realisable? I think, there gives it still right narrow Limits. On the other hand see me, what is Technological on Become? Since is Incredible on Coming. So when one only times everything together, what it already gives. Cordially welcome to the Podcast Good through the Time. The Podcast round at Mediation, Conflict coaching and Organisational consulting. A Podcast from INKOVEMA. I am Sascha Weigel and welcome you to one new Consequence. And today again with mine Colleagues Dr. Frank Termer together. And the means, we are with the new Technologies for Mediators and Consultant.
[0:36]Frank, like goes’s you? Good Tomorrow, Sascha. Yes, me goes’s dazzling. Everything good so far. But yes, the Events roll over itself straight. I white, I am not only Mediator, but also in the Association on the road. And the was for us yesterday, we are today here at the 7. November, was yesterday a whole more exciting Day. In this respect, me goes’s good, that it much go. The shows times again, that it sometimes also to late be can for Mediation.
[0:57]
Exciting developments in mediation
[0:55]Man must already the right Timing catch. But the shall us today not thereof hold, about new Technologies to speak and like them Mediation change and perhaps even better can.
[1:05]And more exciting Day, because the also elsewhere exciting was. The American President is elected been, the German Government has New elections arranged, the Minister of Finance is dismissed been, the Minister of Transport is from the Party resigned. Nope, the is exactly that, what with us in the Association Yes then also arrives and where we us natural also then somehow intensive with employ. We must Yes also, so like it with the Mediation also the Case is, always before the actual Event already ready be, at Things in the Paths to lead. Correspondingly have we us on one Federal election 2025 natural already prepared, the have we rather so in the Autumn thought. And now sees it Yes whole strong afterwards from, that we very, very much earlier on it be become. Since applies it natural now also fast to be and exactly, the right Things in the Paths to lead. Okay. And right Things in the Paths to lead and before the Shaft to come, there are we with our today dialogue guest arrived.
[1:50]
Michael Lardy in the podcast
[1:51]We have today invited, I white even not exactly, whether I now say can, one Technology experts, but on each Case one Mediator colleagues, the early already ChatGPT utilised has and also published, reflected has and in addition one whole interesting Contribution now also composed has. And the was Reason enough to say, Michael, Charge, you must with ins Podcast studio come. And hardly to believe, he has immediately Yes said. Hello Michael.
[2:19]Hello, Michael. Hello, Sascha. Hello, Frank. Yes, many Thanks to for the Invitation
[2:28]
The influence of AI on mediation
[2:26]to the Podcast today. I am very tense. Yes, a mega exciting Topic, AI in the Mediation. So I have December 2022 JGPT discovered and have actually already Relative soon thought, boah, the would be Yes great, when one there Role player so what make could. And the was so the Starting point, that I me then to and to always more intensive so that busy have. We watch us today so to speak also really as Mediators the to. Not so very so to speak from the technical Page probably, but I am also tense, like you so to speak there got there are, Michael, so to speak as Mediator from the Mediator corner or whether you on other Paths with the Technology first times in Contact us come are. And look us but today to, so to speak so one Headline, so like human can itself AI give and like become we so that Bypass can. And there is Yes already some like this, what itself is emerging. Michael, first but to you. You are Mediator, know we both, so Frank and me, and you live in Salzburg already one long Time. Who are you yet? What makes you out? How come you with JetGPT in Contact? I live since about 20 years in Salzburg, am Saarbrücken actually originally.
[3:38]Yes, JetGPT, I was already always very tech-savvy. So originally have I Chemistry studied, had then about 25 years one Solar company, so in one whole other Area. Got then very late discovered, or what means very late, but at least Relative late, discovered, that Mediation actually a fascinating Topic is, at Disputes simple different to solve. And have then, as this Tool came out, thought, the would be Yes really fascinating, when one simple so his Skills simple better improve could, by one there Role-playing games makes, by one Mediation and Simulations makes. And it is stop very fascinating, as I so that started have, could this Tool still extreme little. And the Fascinating on it, find I in any case, I take to, you goes it similar, the extreme Development this Technology. So one must practical for each Date, the one somehow has, a Seminar or a Webinar or a Podcast, one must itself each Paint again view, what can the Part now again everything. The is alone the last Progress with this new Voices, this new Voice function. So the Development is fast-paced and find I absolute fascinating. And I think, it gives diverse Possibilities, like one it insert can.
[5:00]But one must itself also whole good so that employ, so that one it right Realistic estimates. You say, you have the already early so to speak utilised and nevertheless so to speak these Dynamics with into account. So you say, we must us each Paint from New again view, what itself there improved has. Whereupon wait you on most likely still now with AI, where you say, then is the again for us Mediators, for ours Tasks, the we perhaps in the Mediation with Mediants have or the we even also outside, Before that and afterwards manage. You say, when the comes, the Feature, then have we really again a Instrument, where actually all Mediators immediately say must, the belongs to for me now to the Everyday work. Because that, believe me, is still not whole like this, right? Frank, Michael, like says her, is the already so a arrived Instrument? My Impression is, that Mediators but in principle extreme sceptical are, or quite sceptical are, what the Skills from AI and so I speak always from JGPT, because mine Opinion to is it the most advanced Large Language Model in the Just a moment. The notices man, when one so that works. I find, the other AIs come there even not with, apart from of which. I work to the Example also not more about one Keyboard. On most sensible is it stop simple orally, the to make, like it stop in the Mediation last Ends also is. And from therefore, I orientate me on mine other GPT.
[6:24]What there still missing, well, I find, it can already one whole Quantity, and I think, one comes now already hardly behind, so right always to test them out, what it already everything can. I mine, it gives these Studies, that one AI empathetic is. I think, the has some People but right surprised. Mediators generally have, believe me, tendential the Feeling, that it some Skills gives from Mediators, the one AI stop. At all not somehow obtain can. I believe, in the Point must one a
[6:55]
Empathy and AI: a critical view
[6:54]little cautious be. Stay we times with the Empathy, because the is Yes for many Mediators so one whole central Feature, the them believe, bring along to must as Mediators and at the it also goes in the Mediation. And I can me good introduce, when it someone listened to has, the AI is empathetic, have Studies shown, then is immediately, the must one somehow more accurate view. The can actually not be, because the is Yes one Acceptance, that the not and never possible is. What is exactly the Stand at the moment? Also Frank, you are Yes also inside. How would her the Topic Empathy at the moment prescribe?
[7:33]We had Yes here on the Place also already times about the Topic Empathy spoken, when it at ZGPT or Artificial Intelligence in the General goes. And I find it very interesting, also from the Conversations the last Weeks and Months out, there two Things to differentiate. The one is yes, can one actually prove or test or like also always find out, whether so one Technology really empathetic is or Empathy show can. Or the second History, the find I much more exciting, write People this Technology these Feature to. So has one quasi one Kind Credibility on the technological Page created, at People to make or this Faith even to develop, that the Technology actually empathetic be can. And we have Yes also already times about this Tool ELISA spoken, what Yes also Michael in his Essay thematised has, what Yes already in the 60s years then one similar Approach tracked had with wide less more technological Maturity, as we the today have. And nevertheless has one this whole simple Chatbot then already Empathy awarded. And the find I always still times one exciting Point to differentiate. Goes it us rather therefore, really to prove and to show, is these Technology actually empathic?
[8:41]
Humanisation of technology
[8:37]So understands them what? Has them a Understanding from the, what there straight happened? Or rather the other Page, write People so one Technology Empathy to and can the also already for something utilised become? Since would me also, Michael, in yours View times interest.
[8:52]Carries these Differentiation or makes it at all Sense, itself on one from both to focus or should one the somehow holistic see? How is there so yours Impression? So I think, the Example with ELISA is really very good, because then was Yes Weizenbaum, so the Professor on MIT extreme impressed or what means impressed, irritated actually rather, which Skills the People, the ELISA utilised have, this Computer programme, the Yes in the Ratio to today very simple was, what them the attributed to have. And it gives Yes many Examples, that the Humanisation from computers, that the one very human Tendency last Ends is, a very simple Computer system. Take we the Example from the Tamagotchis in the 90s years. So it gives one very beautiful Essay from one American Researcher, Sharon Turtle means the, believe me, the the analysed has. And the was Yes really extreme primitive and nevertheless have the Children the very extreme serious taken, have the lined, had Fear, that it dies and so continue. So the Tendency, Technology to humanise, is, believe me, already very pronounced. But watch we us so to speak these Difference and I find, the is in the Reason taken the communicative Problem.
[9:59]Humanisation on the one Emphasis page, we make the Machine more human, ultimately so to speak the Frankensteinian, we pump Blood pure and breathe him Life a and then would it so to speak one Kind biological Life be, the then also to the Empathy possibly capable is. Or but communicative watch we us the Addressee page to, so we, which Picture make we and what calls it with us out, when a Human, one Cat, a Dog or even a Computer programme and then as Language in us something in Movement brings, us so to speak emotionalise can. And the finds but with us instead of and not in the Machine. And there would I already say, so the Decision is important, but them is not so Important for the Question, what the Machine can, whether them like a Human is or not, but what them with us People make can. And we are, and the believe I is also in the Study, the you addressed have, Michael, whole clearly, we are total fascinated, emotionalised and itself sceptical People can itself the hardly withdraw. When man, and there find me, is the Voice input actually one Huge hurdle, where also then…
[11:16]
The role of language in communication
[11:11]Everyone, the speak can, immediately with the Machine communicate can. And the Chat is Yes but rather again, I would now not say north, but the is so to speak like the Threshold earlier, whether the Interface in the Computer understandable is or whether the what only for Technician is. I believe, there comes but also still what in addition, Sascha, namely the Kind and Wise, like ChatGPT there also reacts. Because the is in all, what it does, always very, very favourably. So no matter, what you there enter and like you speaks in and with which abstruse Ideas you at the Corner come or Other, you will always first times Consent experience, positive Support, Reinforcement.
[11:49]And the is, believe me, what, when something like that with us resonates, in fact, when we someone find, the us supported, the us positive minded is, quasi one similar Opinion perhaps then also developed or us strengthened with the, what we plan. Clear comes the good and positive to. So I believe, the lies then natural also with the Interface, hangs the also together. But it lies, believe me, to the one or to the essential Part on it, like these Machine then also reacts and like them itself expresses on the Jobs. Empathy from JetGPT, what mean you? Straight because the Voice always better becomes. So straight the last Progress, where now in this new Audio model, the now before short published been is, is Yes also to the Example hardly more English Accent inside, hardly more Error inside, when them somehow one Listing makes from firstly, secondly, thirdly. The has them earlier not so good skilfully. So the Voice becomes always better and is meanwhile actually extreme human. Many things leads in addition, that one Certainly there Expertise reindenkt or with pleasure reindenkt. Yes, where one itself superior must, has them them really?
[12:45]I mine, this Test, the I there cited have, this LEAS test, the is a Psychotest, the one stop also with People makes. Since cuts JET-GPT simple above average good from. With the Tests must one however a little watch out, have I now noted. Sometimes are it the Comparison groups and like this. So them become also a little turned in the Direction, that it especially good cuts off. So at the Teaching test to the Example were it Students only, so one whole narrow Target group, now none wide sociological scattered Group. And now gave it Yes before two Weeks a new Paper from DeepMind, the also published have, or respectively the Press has it actually from it made, the AI would be better as the Mediators. In the Final effect were Yes not Mediators the comparison group, but it was one Person, the none Mediator was, the but to the Mediator made was.
[13:33]Man must itself already the Studies very exactly view, like them constructed are and partly are them also a little so constructed, that what very Positive thereby comes out. But nevertheless, I mine, also when it only approximately so empathetic is, I see the always before the Background, that simple the Dynamics the Development Yes incredible is. The means so, it is Yes not the Status Quo, the absolutely counts, but what have we in two months, in three months, in five months, in one half Year elected. And there believe me, yes, I mine, the AI learns last Ends Yes from the literary Texts to the Example also, there is simple the Emotion inside, there are interpersonal Relationships until to the Do not run again, the is the Basis. So from therefore is it stop one simulated Empathy, but the mine Opinion to always better become becomes. And we have this Element of seduction, when the so benevolent is.
[14:26]I can me still remember, I must me on this Point mostly on it remember, on Odyssey in the Space from Stanley Kubrick, the HAL 9000, IBM-docked Name. The has Yes then creepy Things said. in one Peace and quiet and Friendliness can me still on the Shower remember, as he simple the there sent out and the not again pure left and could the not make in all Question. And these Irritation have we still not. With Computer voices today, with ChatGBT are always benevolent and them speak us always somehow very to. But I can me introduce, the changes itself abruptly, when we these Alignment not more would have and the would something Critical say. Is it that, Frank, where you you mean, there lies still a Field of experience, the us so far saves remained is.
[15:10]
Trust in AI: opportunities and risks
[15:10]Yes, so the is quite conceivable, that the still in so one Direction goes, whereby I already also the Provider resp. Manufacturer there very strong perceive, that the there also much readjust and Boundaries move in and also the Models so to speak in the Bridle hold. Straight in the Starting time, where these whole new came out and similar more.
[15:27]Since were Yes the most abstruse Things already tried out. The Human is Yes then equal so and wants then try, such Systems on the Boundaries to bring and about it beyond, what Yes then also folded has. And the Provider, Developer, Manufacturer have Yes then also quite fast also Boundaries installed, whether the certain Reactions fail to materialise or certain Themes tasty are. The was on Twitter, no? Exactly, it gave so Twitter chatbots and Similar, the then within from few minutes to one right-wing radicals Nazi were and him nasty a pair Wohlen somehow issued have.
[15:52]Microsoft bot word was the then been. And something like that has one then natural prevented and the finds also now natural still instead. So I believe already, that the Human there with his ethical, moral, value-based Performances one important Factor current occupies and also the Company, the behind it, this Frame specify, in the itself so one Technology moves, at also quasi one Development in one positive Direction to drive on and the even not in one Corner to bring, the us not pleases on the Position. And nevertheless are we yes, the have we Yes also already in this Podcast discussed, wide, wide away from really so one general artificial Intelligence, the really one Kind Awareness developed and own Opinions developed and own Strategies builds up, what perhaps Yes with HL9000 and the Film, the you addressed have, already a little anticipated was in fact. When he then says, sorry Dave, I can’t do that. The become we then here with Chatschiputti, believe me, first not experience. In this respect remains the Responsibility already still with us People a Piece wide, yes. But there must I short in between go. Before short have I with one Open LLM version one Llama 3.1, because the also a very exciting Topic is, Open LLMs to utilise. Since have I one Instruction given, he shall one Set reformulate. Since went it at the Ability to empathise from AIs.
[17:02]And there have I so one Hall effect had. And although says me the AI, yes, so there is them but not mine Opinion. One AI would be not empathetic, them would be from the People programmed and so more and so continued.
[17:15]Has the Tasks not fulfilled, the I wanted, namely one Set to reformulate. And has meant, instead beat I you but the before. When one more often with AIs works, the goes you probably also from and to like this, one has so Moments, where one trims, where Situations happen, where the AI on one Kind reacts, mostly rather positive, because in the Case was it rather negative, because them simple the Labour denied has, where one itself already wonders, what happens there now last At the end. So whole so hair-free have I the Feeling, are we not absolutely always. Yes, me was the Point there important, because you the also addressed have, Michael. The is, find me, in the Application a A giant step, that the Thing speaks. So, that we one Voice input have and that it one Voice output gives from one Voice, the very human becomes and the also Filler words utilise can etc. And the is what whole other is in the Perception, so for me is the so to speak significant, I white not, like it for you is, therefore goes it straight, as when one the writes in, as when one with the chats, so chat is somehow really technical, so then sees one somehow, one needs so a Pad, one must then type and one must the Language enter, but the Language, the is immediately somehow more natural and the is for me in the Application and Michael, you say Yes also, you speak actually only still and hardly still so to speak in the Chat, that the genuine a Booster is for these AI development.
[18:40]
The future of AI in mediation
[18:41]Yes, definitely. So I feel the extreme like this, because it is, when one now Mediation applications makes, any Communication exercises and like this, then is it simple much closer on the Reality. I think, itself when one type would, I am now not so fast in the Typing, but I believe, itself in the Typing, it is simple, our Tool is the Language and the Speak. And then comes now also one Language back. The simple incredible good is in the meantime, yes, find I also