Learning and Unlearning with AI: Skills That Outlast Tools - Graziela Tonin

0:00A recent study uh from N MAIT uh shows that uh they analyzed two groups of students one that used GPT and other that didn't use and this group that used ChatGPT reduce uh 47% of their cognitive uh capacity. Uh another uh research shows that 90 95% of the companies that invest in eye didn't see return. Just five percent C value or can measure the value. Need to step back in some some ways and rethink what kind of word need you delivery or create and for who and how because w you know At least for now, this is our decision.

0:57The environment don't need us. We need the environment. The environment could survive without you. Welcome to Software Testing Unleashed, the podcast for testers, test managers, and developers who want to create great software. My name is Richie, I'm software quality coach, keynote speaker, and author, and today I have an episode from the OOP. Conference in Munich for you. My guest today is Graziela Simone Tonin. And we talked about how what we have to learn and what we have to unlearn these days in the age of AI.

1:29What should we learn and unlearn as a individual, as a organization and as teams. Enjoy Graziela's perspective and have fun with the episode Hi Graziella, great to have you on the show here. Hi, thank you. Thank you for inviting me. Thank you. Yeah, it's it's a pleasure for me. Uh we have now we are now here in the OOP in Munich 2026. the second conference days sliding out uh last talks are running and uh do you have uh your talk already Yes, yes I did yesterday.

2:11Ah, okay, so you're now totally relaxed when you're here in the podcast. I read the abstracts and I saw uh your your speeches about uh how to deal with all learning and also unlearning uh in in the age of of AI uh nowadays and for organizations I thought oh that's a very interesting topic I want to talk about. Good. Thank you. You know, uh nowadays the question is no more what a a I can do it or what a I can do a I can do better code. Um Sometimes even strategy, write tests, and any any task, you know.

2:56So the question is how we can transform or unlearn uh institution as universities, business, how will be what will be the unlike The unique human value in this in this new world or new new era that we call the compute supercycle or s compute AI co uh supercycle, you know. So the idea was Bring some discussion data and also recent research that show how the society is trying to face this challenge, you know, because As a professor we completely need we need to completely transform the way we teach, the way we use methodology to provide knowledge for the students, you know?

3:54But also has a business. The business model is transformed completely transforming. And sometimes it's not just about uh specifically when we talk about human habilities maybe or kompetens. Not just about how I can do a course and understand AI or these LLMs or this new you know pro how to do a profit. But it's how I create this new and drive this new world and new era as a human leading it, you know? Yeah, yeah So you you mentioned uh the the universities and and also in the abstract there was this the school before university.

4:34Wha what do you think? What what could be a change there to reflect the modern world of And what do we have to learn där för den för den world nu i förtär? This is en hör question. But deting is. The young generation all of Denis using AI to study to do uh College tasks and even at university do the tasks they need to have homework and so on. So we have no choice, you know? We need to understand how we can create safe environments as in in university but also in business.

5:19and do experimentation and try to understand how we can use it to in fact learn, evolve, or has a tool or has a co-author implementing software and not just author, you know Because when um a recent study uh from NMAIT uh shows that uh they analyze two groups of students, one that use GPT and other that didn't use and this group that used Chat GPT reduced uh forty seven percent of their cognitive uh capacity you know they didn't remember after some days they completely didn't remember the context they didn't uh learn in fact So we are missing uh cognitive uh abilities, and and this is important, you know, it's not just about develop a new competence that understand a new technology and apply it to solve complex problems but it's about didn't Missed RKI, you know, because it's a challenge.

6:32And we have some good experimentation that some university companies are do are doing to try to uh in increment it or implement it in uh universities such as definitely university need to transform the model to teach and share knowledge. The professor will not be more the holder of the knowledge. Was the university will be not the the place we go to get uh information knowledge we are surrounded about a lot of information from many many countries many many good universities in any place we are you know we can access it through internet.

7:16So the idea is probably University will be a hubby where we can um Know that the the content they are sharing is a trustful content university design you know design the journey of my professional career helping me to make good decisions helping me to to drive um this journey but also help me uh to develop critical thinking, ethical issues, you know, think about that, question about that. So the professor will be more a guide, a sponsor that support my process of learning, my own journey.

8:00And for sure, AI could help to personalize it. For example, in software engineering in computer science we have many many examples when we use AI to tutoring and they are there with the students 24 hours professor uh No and also they could support each student in each level personalized for that student and that challenges the student is facing during that process or that exercise. Even if it's simple exercise or complex exercise The A I could help to improve my knowledge in this sense or in this way.

8:41And university could help to develop these tools. and help students understand how to deal with it, how to use it in a good way that helped me to have a superpower or super intelligence instead of drop my my my reduce my intelligence And but you know uh the thing is we are we are seeing that most of students are using uh a cloud. In computer size subject to implement code. A Claude's uh research recently showed that The bad thing is that people are just asking for AI to delivering for AI the task, you know.

9:25AI do the task. They are straightforward asking the task It's not about co-out or or discussing or collaborate as a pattern, a partner. But they just asking and and transferred this responsibility. This is the risk and it's it's um a problem we are facing now you know. So we have to also to learn to to use uh the AI stuff more appropriate so that we don't uh don't just put in what we want but to to make a process out of the working with the AI. Yes.

10:00Because uh for sure we don't know the best way to use AI. We are just starting with many models and uh uh uh now generative ai is fails and everything the hype is just happen and we we are starting using a lot of agent ai that the power is super power that could make better decision sometimes than uh better than us because can consider a lot of data and and data and so on uh but If if you didn't collaborate and try and share these initiatives to understand how we can um Develop this process frameworks and waste to to learn and unlearn.

10:54I can't believe that it would be possible we we create a better environment because you know things happen in a unprecedented velocity nowadays. And if for example, one one thing that I really believe that will be important is sharing knowledge community. Has it say in Agile, Squad uh or squad methodology to things like um uh the guildas we have to to to discuss, to collaborate in some areas and and create better solutions to complex world. So in one side we have a a world with a lot of complex problems, geopolitical problems, war, uh environmental issues and so on.

11:41On the other hand, we have like a lot of good technologies and knowledge. uh that we can use but changing fast in unprecedented velocity. So in my opinion if you if your university will be a hub if you create more communities to sharing this uh knowledges and try for example you are a specialist in quality in test and so we can together uh discuss your knowledge experimentation and our um learning in this process Maybe we can uh face or uh discover better way to deal with it, you know?

12:21Has has was for example I really like to use this example in the pandemic time uh the vaccine was possible in a faster way to to be developed because this col huge collaboration around the world with many scientists discover a piece of research and connect And it connected uh results in a vaccine that saves a lot of people, you know. So for me, this It means that probably we need to collaborate more as a human. Yeah. Yeah, and that's a good good way of learning too.

12:56Yeah. So uh you mentioned uh also the the word of unlearning things. Uh I I think that very interesting because uh Well what what do you uh in specific mean? What what should we unlearn? What what's behind that? My vision about that is that uh because as a human, for example, uh Agile principles starting to emphasize that we need to think more about sustainable work, sustainable journey of work. Be more empathic, um, collaborate instead of just competition, deliver value, and these kind of things And I believe that we forgot or we miss this kind of essentially human qualities.

13:45and abilities are our are behaviors. And we we we just aren't in uh like a robot sometimes, you know, acting, facing, challenging, working every day a lot. Uh we we saw that increased the number of problems, main mental health problems in companies and in fact it reduced our It's a correlation, reduce productivity, reduce result for the business that is also not good. But We miss these kind of things in my opinion, you know. We just starting trying to make money in delivering new products, new services, and Be in this cycle every day, every day, every day, more robotic in way of work and delivery tasks and so on and automate things, but instead of considering the impact, uh considering the environment, the the people, the society, and also the uh the inequalities, you know, because we have a lot of um challenges.

14:46Yeah. Uh that we not consider we say that for example in Brazil we have like a small city that they don't have still don't have electricity. But they will provide the city, the ecosystem, the government, the the private sector will invest millions and millions of dollars in a new um in in a new infrastructure IT infrastructure to cloud to you know to provide uh to provide space to ai to uh to to have to save data, you know, and and and improve the models, for example, you know, data centers and so on.

15:26Data centers and so on. So uh it's a paradox. And uh for me on learning is something like that. You need to step back in some some ways and rethink what kind of word you need to deliver or create. And for who and and how because w you know did it At least for now, this is our decision. We can make this decision. We are guiding the AI I know that it there is a monopolization of the AI models that is something we must concern and we cannot control maybe.

16:10Is a human guiding and driving these changes? So maybe we need to rethink a little bit and and and learn something, you know, how to do uh a work uh for example, you know. Yeah, yeah, it's a good good inspiration. So w when we transfer to the to the business environment, to companies and organizations, what do you think are the uh when we when we look at the people there, what what is their ability to learn? What should they learn now and what they have to unlearn?

16:41Is this this different or do you have some Um in some way it is because but you know it's complementary but when we are doing business nowadays and i in many many companies that I'm visiting or or lead a project I still saw like a top-down leadership where the sea level or or the high level make the decision, define the strategy, and so on, for example. Just one example And imagine now with all these challenges, all the complex problems we need to deliver a solution, to improve our business and also to stay competitive in the market and because the time to market.

17:25Doesn't make sense to have this kind of leadership nowadays in my opinion, you know? We need for example change the leadership way uh in the companies how how we can uh make decision, how we can define strategies such as I saw a lot of companies uh another uh research shows that 95% of the companies that invest in I didn't see return. Just five percent see value or can measure the value, some of them maybe even measure something or could measure something and see something.

18:03And why? Because they are just hyper-AI driving. And not like a concern decision, estrategical decision, um connected in all layers of the companies, not just uh isolated uh experimentation that they try to do something someone there try to do something to help the company you know so this kind of context that agile Suggest something like that. You know, we need to collaborate, everyone to make better decisions. First to define it as a strategy. AI will be a strategy.

18:40Must be on the governance because also for of the risk and the challenge, you know, cyber security and so on. Um and to do that it's important that the leadership guides This culture. We cannot assume that we'll have it without a culture guided by decision made by data. Collaborative discussion, safe environment, so I can test, try, experiment, and suggest better ways to do something Because sometimes I'm in contact with customers. Many times people ask me as a consultant in Agile how I can save solve it I ask then you are in contact with the customer.

19:26What they tell you? I don't know. I'm not you are 20 years in contact with customer. You listen then, you know, let let's start listening then maybe we can make better decision and collect data and consider this data. So in many layers We need to rethink a little bit the organogram of the companies, how leader act. Because leader will not be like Most of the time we consider leaders someone that are just defined the strategy but did in fact lead and make the architecture of this strategy, uh with defining the priorities in each department, in each situation, discussing with the leaders of this department, for example.

20:08We need to redefine how to evaluate For example in software. If I consider velocity, I never trust that this is a bad a good KPI, but if you consider velocity has a good KPI You know, everyone can deliver a lot of things now with AI and maybe destroy the security that the that you know uh input a lot of vulnerabilities and flaws in the code, you know, so doesn't make sense. So we need also relearn how we can uh define KPIs to achieve the results of the companies connected with the strategy in in another way, in a new way of collaborating as a workers, you know and uh the the World Economic Forum show our research that at least uh forty percent of population will need to Re-skilling or up skilling.

21:09That means learn a new skill for the future of work, future of job. or at least uh unlearn or or or transform the skill you have to adapt to the new needs of the market and and and this era I I believe that is the the first the main things you need and also this this idea of that alone as in your company we can go fast or go forward Probably you need to think more about collaborate with m more partnerships, collaborate with universities, you need to bring this this gap, you know, uh create a bridge to reduce this gap because We need to deliver for the market as a university a good professional, you know, and we are like responsible for this uh development.

22:08But for example, if you consider software engineering, we are saying that AI will uh drop the the junior software engineering job. Yeah If it I don't have a junior software engineering, how I can have a senior software engineering to make better, you know. So I know that maybe I can do better work, but so how I can solve it? And my suggestion is improve the connection um with universities in the early days of the the students at the undergraduate uh course uh sharing so uh in software expert software engineer sharing the impact of using ai in a better way the risk, sharing how to deal with complex problems, how to solve real problems, bring some problems to students test, um bring them to to the the real world and connecting then so we can use more PBL approach to to teach that is problem based or project based learning So this connection is so important to and also it it helps companies um connect the young students with their culture.

23:24starting um teaching then early the culture of the company and probably when they go to the market with more maturity they will be prepared or better prepared to Go inside your company and be part of your team and create and deliver better results for your customer, for example, you know? Yeah, yeah So like we have some challenges. Yeah, there are some such some challenges. For for us as a software engineer, tester, developer or something, do you have one one tip How how what what skills should I look for my own?

24:00What what what should I more learn or what in in in a personal way? We are discussing a lot about how to improve curriculum nowadays, for example. What kind of competences would you like to teach for the future? uh professional so one thing for me is learn how to learn we call that you know because we have you know yesterday I was looking many many many data about AI and the educational revolution. But which one I can trust? Which one is the best data?

24:34You spend a lot of time trying to understand the amount of data you You you can access even you could use AI to bring you some ideas, but you need to trust in that data. So learn to learn um how to solve complex problems. Because you know One thing that is for me in my opinion is unique human is okay I have a customer they that have a complex problem But it's in that area that I know, but always they have a specific characteristic is that is that customer.

25:11And you go there and you need a specific way to negotiate, to define the requirements, to define the schedule, and so on. We have a lot of opportunities in our area, but you need a good team working. So know how to work in team it's important because it's impossible to deliver a good and powerful software alone you need at some point you you could start alone but at some point you need someone uh to help you to scale and and achieve more customers and so on Critical thinking, how learn to learn, uh ethical, it's so important for me.

25:55Today if you ask me if I will hire Uh analyze some curriculum to work with me for me commitment is the most important thing, you know If the person is ethical, personal, could behave or respect diversity, because how to learn a tool, I have many tutorials and materials that you can say, oh I can help you. And we learn together this two, you know? And tools change, languages change, frameworks change. But The way you think, you behave in front of your customer, you you you treat people this is so important even to uh to to gain contract to to make money you know So this kind of ability in my opinion will be so so important.

26:46Because for knowledge in any areas, we will have a lot of material and support that help us to understand that That contains in learn about you know. It's very inspiring because uh yeah that gives us an an idea of what we can learn and to to face all these challenges we have we have in the future you know with AI now. A and and to unlearn. Yeah. Just to to answer and maybe connecting with I said is how we think about we make business how we think about we define uh our career or how the amount of time we spend with Things that matter.

27:34Yeah. Like with families or relatives or so on, you know. And even for us, because this this recent research that showed that uh mental health is a issue for many many companies and it is increasing uh through the time is because we didn't save time for us. You know, to do something that you like, go to a park, I love play piano, uh do yoga. I don't know what is Good for you, you know. So maybe it's important to unlearn how we can connect with ourselves You know?

28:06That's that's a I think a very interesting. You know? And also for sure we need to think more about how to impact the environment. Because a friend of mine of sustainable subject told me We don't the environment don't need us. We need the environment. Environment could survive without you. I think this the thinking about an environment uh Steps a little bit back in the last two years because of all the AI stuff and the geopolitical problems away.

28:39But it's so important because it's the The future environment that we have just this one. Graziela, thank you very much for all these insights. can now think about how to deal with our teams, with our with ourselves, uh what to learn, what to what to unlearn. I think that's a very, very nice uh claim, I think. Um yeah, so thank you very much that you were here. I wish you a great rest of the conference here and a safe travel home and hope you see you soon.

29:11Thank you. My pleasure. Thank you so much for this opportunity. Thank you. Bye bye.