Pioneer Knowledge Services welcomes you to the next because you need to know. I'm your host, Edwin k Morris. I serve as president and founder of this organization, and we are thrilled to offer this educational program. These design conversations bring you people's experiences from all over the globe in the field of knowledge management, nonprofit work, and innovation. Assalamu alaikum. Bonjour tout lemons. Salut. Hello, everyone. Good morning, and welcome to this podcast. My name is Rashad Najjar, and I live in Grenoble, France. The most interesting thing near me is the LHC, Large Hadron Collider. It is the world's largest and highest energy particle collider. So this is somehow scary and somehow a little bit good. So I'm located on the French Swiss border. I love to meet new people. That's why I'm here on the podcast. I'd like to connect, to share my stories, listen, to their stories, and offer help if they need it. The things that I have learned, throughout all these years is that preparation, persistence, and patience are keys to success. Half of the success, I believe, is preparation and determination. The topics that I'd like to talk for hours and hours is knowledge management, obviously. And I recently wrote and published a report exactly analyzing generative AI use cases on where and how those capabilities can augment the learning experience, and the knowledge management processes, of course. So this has been my career, 10 years in this field, organizational learning and knowledge management. The last book that I read that I'd like to share it with you, is really very interesting book. It's called the big on by Mariana Mazzucato. She's a professor in innovation economics at UCL. Mainly, the authors, are arguing the myth that the consulting industries always add value to the economy, And they support this argument or this counterargument very brilliantly. They said innovation should come from within, from the inside out, investing in the collective intelligence of the people and empowering them to pop up ideas and to build on the collective elaboration of those ideas and do innovation. Is AI gonna go away? Is artificial intelligence just a fad? It's been here since 1983. It will continue to be. But like any other technology, there's a hype curve. It will go and up based on the drivers, the technology, overall circumstances. So let's demystify what AI is. Yes. I think there's a lot of folks that hear it and they see it everywhere now on every social media platform or every entry point they have to news and media. What is it? Why should they care? Perfect. And, I like to demystify it. There's a misconception. AI, artificial intelligence, is intelligent. This is a misconception. It's not intelligent. We are really at the extreme computing edges. So we have very powerful computing machines that are doing enormous computing. Other hand, let me ask you this question. True. Can a machine wake up in the morning and says, what should I do today or what I should learn today? Well, that's self consciousness. Right? That's a self driven reality of purpose and Yes. Maybe someday we'll get there. But right now, you're saying that ability is not what artificial intelligence is. Yes. And most risky thing is our expectations. So we are expecting from AI much more Yeah. Than AI is capable to do. So just just so we can add more clarity to this, so I looked up Merriam Webster's definition of intelligence. Yes. So artificial intelligence, as you say, is a kind of a misnomer because it's not really that yet. Yes. Intelligence in the first definition is the ability to learn or understand or deal with new or trying situations. I apparently have never read that definition before because that's pretty cool. So that is definitely a human trait, for some humans, I'll say. But that is a humanistic type trait. You know, let me share also with you something. I did a PhD studies. And during research, we are always and most of the situations, we are confronted in ambiguous situations. We don't know what to do. And there's no reasoning or there's no, let's say, evidence if we are going to choose this correctly or not. The only thing that drive us toward a direction is our gut, something inside us. Does the machine have a gut? I'm just questioning. I don't know the answer. I'm questioning. That goes back to the 1800 with the original concept of the thinking machine. Yes. That this idea that we can mechanize comprehension and understanding has been around a long time. Definitely. And let me give you one more thing. There's shared GPT gains the mainstream attention. But do you know that, the training because, you know, AI is a subfield of machine learning, and ChargeGPT is a subfield of deep machine learning. And deep machine learning, there's 2 phases. The first phase is the training. This is where the machine learn. And the second phase is the fine tuning when we tune the machine to adapt to the outcomes so we have an accurate outcome. For example, if we are applying this, AI in the field of cosmetics, So we need to teach the machine the specific lexicon or the specific knowledge about cosmetics. What I'd like to share, chat GPT and all other large language models, they use something called ghost army. So the ghost army, those are workers who do very millennial tasks like labeling an object. This is a car. This is a fish. They say, yes. This is a car. So the machine doesn't know what's true, what's this is a ghost worker, and they are paid $2 to $9 per hour. And they used a massive army of ghost workers to train or or to validate the outcome of the machine and to say, this is true or not or to label the objects. So now how can a machine be intelligent and we are still Right. You're still in need of the human perception and the human expertise. Well, you're the way you just explained it, it can't be done. Unless, we get to the bioengineered level of wetware, which we can program hardware software by our own synaptics structure. Maybe. Yes. Then you've got embedded humanity in a machine. Well, this is a call for all our entrepreneurs to take this idea. The Rise of the Machines is the article I'm looking at right now from Smithsonian Libraries. Yes. And we talk about how this whole idea that spurred off the industrial revolution became this idea of self manipulating machinery. Yes. And it's just bizarre to think that this was back this these concepts go back to the 1800 that people have been trying to make machines, so called think, but at least make a cognition differential in their mechanical approach to whatever the job was. And a lot of these were very labor intensive work for these machines. So if we go forward Yes. So this is 1800s that this whole idea of thinking machines, and that's just the general terminology here. 1800s, thinking machines. 2000 twenties, AI is everything, but it's not really intelligent. What's a 100 years from now, my friend? Well, I wish I would know. You're a pretty smart fella, so I think you can conjecture, please. Yeah. I think we will continue to advance in computing. That's true. Previously, during the 19 eighties, 19 nineties, Moore's Law is, his Intel CEO. He said the number of transistors that we can put on the same area, fixed area, will doubles will doubles every 2 years. However, what we saw with the Gen AI, since 2013, the computing power has doubled every 3 to 4 months. Yes. Definitely, when computing power increases, it will unlock for us more a lot of, new capabilities. And what I can foresee is really still within the realm of pattern recognition for complex pattern recognition. For example, if we are studying the drug effect, if a drug if I am on a treatment and I am taking a a drug and I have my personal health record and personal health characteristics, am I on a diet or not, and what are other drugs I'm taking? So this kind of interactions between the drugs or my personal health records, those are very complex patterns. So AI here can help in extracting or identifying relationships between all those different variables and recommend the best scenario and the best pattern to follow so we can have, I have can have a treatment without any kind of Mhmm. Side effects. Well, that's a great use case, and I think that shows the value pretty easily to the majority of people Yes. Because it's something we all have to deal with. So it's good to know that that use case is in the positive. But is there a negative side to it? Oh. Is can it be criminal yeah. Okay. Alright. Alright. Let's let's hear about the negative side to all of that. Yes. Well, I can think of 2, negative consequences. The first one is about, security, privacy Mhmm. And personal data. How much I am willing to open up my personal health record, which are highly confidential information. So maybe some commercial companies will leverage those data and build some commercial activities or some or make some profit. So this is very critical topics related to ethics. How we can build ethical AI applications that are not harmful, that respect my personal privacy without being judgmental or being intrusive into my life. This is very big and vast topic and need the involvement of multiple stakeholders, multiple perspective from legal, from society, from health personnel, and every single decision maker. So is this a call to action? Do you see this as a call to action right now that somebody on a unified global front get some teeth in this? Because right now, I would suppose most countries at different levels are all trying to combat this in their own way, in their own domain. But this is a global issue. Global issue. Correct. And we should learn from the past from specifically the GDPR, Global Data Protection Regulation. Mhmm. We at global level, we didn't took it seriously until it was imposed and until it was confirmed. And then companies rushed to comply with GDPR data protection. And the the European Councils has made some law enforcement. So we should anticipate, this trend and work together in advance to put some common, let's say, principles and values, to deploy and develop ethical AIs. What I'd like also to mention that ethical AI should not be confused with a checklist. We don't want to create a checklist and say, yes. We comply with being transparent and being respectful. It's not about checklist. Ethical AI is the process. It's not the end product of the process. Every single application, every single time we work together and develop AI ethical, we should consider it a case study, and we should reevaluate that application with respect to the principles and values that we deployed. For example, face recognition. I am a company, and I am developing face recognition application. So I should go to the checklist and say, yes, it will respect my private life or not. Instead, what should I say? Should I evaluate it as a case? Who are my target audience? Should the public be able to use this, face recognition, or I should limit it to the celebrities? Mhmm. Or should the government agencies should do it? And if the government agencies Or or who's got deep pockets? Right? I mean Deep pockets. Yes. And even if the governmental agencies are allowed to use it, are they allowed to use my personal data? And what type of biases I should eliminate? Because in when doing face recognition, there's a skin tone. The dark, the yellow, and different skin tones. So I should my training samples should be representatives, so I should remove biases. And my applications doesn't give me or favor white skins. Yeah. Yeah. Dark skins. Also should be considered and reevaluated carefully before deciding or before opening up this application. Let's take that a notch further. So I'm still looking at the GDPR. So this is a year Yes. EU, European Union Yes. Product. And I wanna get back to this, so don't let me forget. But what you're talking about relates to me in my history as a broadcaster. The Federal Communications Commission in the United States was built to help protect people from media. That's my generalization of the FCC. As a FM radio broadcaster, we had regulations and stipulations on what could be said on the radio to protect the public. Right? But those regulations were regionally constructed. Yes. What plays well in New York City will not play well in Idaho. You know? So there's regionality and sensitivities involved. So does that call into this conversation, the idea of having a regional gatekeeper, a regional meaning that there's an interpretive level of the overarching controls, the data policies and enforcement, to allow for pockets and regions and cultures to have their own flavor. Don't get me into politics. Oh, that's not politics. Come on. What sparked that was that, as I read the GDPR information on the website Yes. The history of the GDPR started in 1950 with the European Convention on Human Rights. Yes. So this has a deeper level than just data protection or cyber anything. Yes. It's just human rights. Yes. Correct. And it should be globalized. But what is the risk? The risk is then we customize. We say this is too generic, and let's customize it to our own preferences. And there's a much more customization for every company. Every company will say, yeah. Those are good guidelines, but I have my own values. I have my own mission. Let me customize it. And then slowly, what will happen, we deviate. And then when, diversions will happen when we have a little of deviations, and then we go into all directions. We need to take it seriously and work it at global level. So is that a UN charter? Does is the UN involved in this? I I wouldn't recommend the UN, but, Oh, okay. Alright. So who should own this? It should be community owned because involving end users because it's all about being transparent and being truthful to the end users, to the customers, to the users. So the users and the customer should be involved. And in fact, there's something also that I'd like to highlight. Gen AI or machines, they don't make decisions. They are those who the developers and the designers and the programmers, they ingest their own, algorithms Mhmm. To make decisions. And this is the phase where the customers need to be engaged with the developers and the designers as much as possible and explain to them how the machine is making decisions. For example, if we are taking, the radio imagery in the medical field. If we are developing an application for imagery and MRI, we should design it in a way that it will help the doctor to do better his job, not to replace the doctor. And the doctor should understand how the machine is making decisions and making the suggestions. And so we need to to develop it in a white box approach rather than a black box. So who owns it? Who who does all this? It is community owned. We need multiple perspectives, everyone to be involved, and it is a continuous process. It's not let's do it and forget it. Develop it and deploy it. It is a continuous process. So I'm trying to think of the software development or web development. There was an organization out of California. They did the protocols for HTTPS or or I can't remember what what their construction was that aided to web interfaces. But it was just a bunch of concerned citizens working together to make it happen. You know, it was an open source Yep. Kind of reality. Yep. So are we looking that that's what we need so it's not part of any government? It's not part of any corporate agenda? It is just the people with human rights that want it. Yes. And I am in favor of a decentralized model and distributed models, which is which lead to the open source community. And I advocate this kind this is working mode. Okay. So when when are you gonna start that? Let's do it tomorrow. Alright. Alright. Let's go. I'm for it because I don't see anybody owning it. And if anybody does stand up to own it, they're only gonna represent their little piece of the pie. This has really long and short term consequences. Yep. Yep. And, also, we should not forget the environmental impact. So this is more related to compliance and ethics. But, also, there's second negativity back to your initial question. To share with you some facts, let me bring up my cheat sheet. K. Okay. So to give you some facts, you know, the GPT 3, the 3rd GPT 3 Mhmm. In terms of water consumption, every 25 to 50 questions, we ask to GPT 3, the machine or, let's say, the whole cluster consumes, one bottle of water every 25 questions on average. The water consumption is 1 bottle of water, around 1 liter. And the c o two emissions emitted by the cluster to train GPT 3 is equivalent to 60 flights between London and New York back and forth. And the energy consumption, because those are really, hungry clusters, they consume the for GPT 3, which is 1,000 times less powerful than GPT 4, it consumed the power or the electricity equivalent to 90 households for 1 year. So this is a real environmental impact in terms of water, power, and CO 2 emission. And those are just to for GPT 3. And now we are talking about GPT 45. And Microsoft, as they just published their environmental and sustainability report, their water consumption spiked around 25% year over year from 2002 to to 21 to 22 to 23. So every year, 25% increase in water consumption. So this is a really a serious issue or consequence to consider. Those clusters used to train Gen AI are hungry, are thirsty, and they are they pollute. So we're back to the birthplace of a lot of pollution in the globe to industrial revolutions that have taken place that had huge consequences. So in parallel to, like, coal and other industries that made everything polluted. So now we're at the point where people that think, oh, it's on the Internet. Oh, it's it's somewhere. It it just happens. It's all magic in this little box here. Mhmm. There's no connection to any consequence and or output requirements or input requirements to the process because it's not visual. There's no I'm not shoveling another shovel of coal in this machine to make this work. Yes. It's just all magic. It's just it just happens. So who should be looking at that? Because this is a real world situation that is building up and building up as it it becomes more user friendly and more user interface is happening, then that consumption keeps going up. When will it stop and when will it actually hurt to where it's news? Where it's actually covered on the news. Oh, this this whole farm underground that feeds this machine has consumed x amount of water. Now the state of Nevada doesn't have any water. You know, when is that gonna happen where it becomes a critical juncture that things have to change? Well, I'm not trying to be a pessimist, but already we have the global warming we are facing. And how much actions we are seriously taking to to stop the global warming. Yeah. And we are seeing all the natural disasters from, the record temperature to the flood everywhere, and we are seeing the consequences, the natural crisis. So we don't want to add a new layer, an extra layer on top of that. Right. And we should try to start with education. Universities have a critical role When we are teaching Java or Python at universities, we should add a course called programming for sustainability or developers sustainable developers. And we should integrate the notion of environmental impact into our coding and programming applications. So we should start doing the educations. Companies, they have their corporate social responsibility, affinity groups, local communities. Everyone should tackle it from his own perspective. And all of us together, we should take it seriously and to be energy effective and to try also to implement to do some kind of decarbonization. I know from my company, we are launching a lot of programs into being carbon neutral. And, also, being carbon neutral will help to compensate all the energy and the all the CO two emissions, emitted from those clusters. So if we put this in the context of every other Yes. Advance that created pollution, I just search artificial intelligent pollution because really we're talking about the the environmental impact in total. Yes. Most of what I see is how AI is helping pollution. So it's kinda funny that most of what I see written is all about, oh, it's all great. But there is one article. Mhmm. And this is from The Guardian. So the source is The Guardian. As the AI industry booms, what toll will it take on the environment? Mhmm. And it really looks like they're digging into what does this all mean? So they talk about Amazon data center in, Manassas, Virginia. I think I just wanna plant enough seeds out there that people really need to start Mhmm. Realizing the world's impact on this rapid access to information and and data. Yeah. Just because it is silicon, at the end, also electronic components are silicon based. It doesn't mean they are clean and they are they don't pollute. So going back on your water, I just see this one paragraph Yes. From the same article from The Guardian. Google became the first this is quote. Google became the 1st tech giant to publicize its water usage worldwide, but provided average figures that concealed important details about the local impacts of its data centers. After a protracted legal battle with the Oregon City of Dallies, Dalles, Dales, d a l l e s, in Oregon, released data showing that Google data centers use a quarter of the town's water supply. Wow. Huge. This is huge. And I gotta say, I I've been in the world of tech and knowledge management as, not as, as kind of an advancer, I'll say. And you're the 2nd person to be on the show that talked about the cost of doing all this AI stuff or the cost of data centers. Let's just say that AI is not the the main culprit right here. Right? I mean Right. Data centers are data centers, and they're a big footprint. They're a big energy sector, a big environmental impact. And, really, that's the bigger story here is that AI is just another user of a data center Correct. That is the part of the problem. So it's not we're not calling out AI as the main culprit here. Correct. Yes. It is a technology that is Yeah. Heavy, power consuming. Well, you've given me a lot to think about. As I go away with my brain in, fast mode, let me ask you this. What's your definition of knowledge management? K. Great. Thank you. That's a great question. I you know, I'm buffering. I'm trying to say some compliments. Yeah. Buffering exactly. It's off you. But, seriously, let me share with you this story. Last week, I was contacted by a headhunter, a Kilometers Post. And while discussing with him, he told me that also he hired another person, another Kilometers manager for a client. And I asked him, okay. What type of missions you are doing? And he replied, the my client was a banking industry, the finance industry, and I needed a knowledge manager to, unify the customer support systems because the bank had multiple branches and multiple functions, and they all uses multiple and different customer support systems. And he hired a knowledge manager to unify all those systems into 1 system. And I told him, really? Is that how you define knowledge management? This is called information systems, not knowledge management. And the way I define and understand knowledge management from day 1 when I started this career, it's all about asset knowledge. 90% of our knowledge is tested inside our head gained from experience, from practice, from solving problems, from being practitioners. And the job of knowledge manager is to help, first of all, to create a space for, the people, the the engineers, the employees, to connect together in a safe environment that will enable them to interact and to create dialogues and conversations. And through dialogues, conversations, they will externalize, and they will extract their tacit knowledge. And the role of knowledge manager is to capture those tacit knowledge that were externalized, formalize them, and make them available to the whole organization. He's a facilitator. He's a connector of people trying to match people together in a safe environment where we are not persecuted or intimidated because we are thinking out loud, and we are just saying our opinions and our ideas, how I define knowledge management in terms of, facilitation interaction, getting the people together to exchange their task of knowledge. So you're looking at it more as a verb. Knowledge management is a verb because it's around an action orientation, And I'll use the word, you didn't say this, but as a broker. You're a broker to the degree where you see where things should go or you connect dots. You extract. As you said, you try to facilitate a better work environment by developing knowledge for everybody. Exactly. And, also, I belong to, a philosophical school where it is specifically, the constructivism epistemology, where they consider knowledge management as a social activity, and knowledge is co constructed together when people interact together. So, fundamentally, it's a social activity. When people interact together, we can co construct new knowledge. I went to Indeed career guide. So for those in the field of trying to find a job, I'm sure you know what Indeed is. Yes. So Indeed career guide states, information management is the collection, storage, management, and maintenance of data and other types of information. Mhmm. It involves the gathering, dissemination, archiving, and destruction Yes. Of information in all its forms. Information management covers the procedures and guidelines organizations adopt to manage and communicate information among different individuals, departments, and stakeholders. So going back to your guy that was talking to you about a job, your viewpoint of what they described that job was was exactly Yep. Information management, and you called it on him. So what was his response? You know? Or her response? Yes. Ironically, the person considered that I don't understand knowledge management, and the person didn't want to, to pursue his recruitment. I said, that's a win win for both of us. Yeah. Right. Thanks. Alright. Yeah. You know, it kinda shows the light on the idea that a recruiter knows everything Yep. That they're shopping for. Right? The second paragraph I wanna wind up on this information management because it gets more to your point. Information management focuses on the level of control an organization has over the information it produces. It requires building dedicated information management systems designed to help the company use its resources. There are nowhere in that description was any people involved. There's no people in this. You know? It's just machines and software. Software. And let me recall the famous HP code. If only HP, news what HP knows, we would be 3 times more productive. Well, I think you could replace HP with any other organization's name, and it's the same. The same. Well, thank you, sir. It has been a blast. My pleasure. I can't wait to have it on again to talk about what's new. My pleasure, Edwin, and I am was so glad to have this podcast with you. Any other great words of wisdom that have bubbled up that you think you need to share? It's all about being generous. What I have learned in those, 10 years, knowledge sharing, knowledge management is about being generous. Not only was the way I share my knowledge or I offer help, but even in my private life, I'm generous with my family, with my friends, with my people. Most probably, I'll be generous enough to share my knowledge, to help people, and to say, yeah. I'm here for you to help. How I can do it? What I can do to support you? So are you saying that if anybody out there is hiring for knowledge management skills, that should be the only question they have on the resume or the interview is give me some examples of how you've shared or been generous in your life. Is that is that really out skills and aptitude, we can figure that out later, but we wanna know at the nucleus of who you are, is this part of you? If you invite me to Chick Fil A, you are the great candidate to be a knowledge manager. So I hear empathy in there. There's gotta be an empathetic part of the human framework that is always engaged. It's more emotional intelligence. Mhmm. We have so many bright and clever people, but we need more emotional intelligence, social relationships. Is that what we need? Behaviors. It's all about behaviors. So does AI have emotional intelligence? Let us check. Let us ask AI and tell them, do you have emotions? Alright, my friend. Alright. Happy days. Thank you. Thank you. Thank you for joining this extraordinary journey, and we hope the experiences gained add value to you and yours. See you next time at because you need to know. If you'd like to contact us, please email byntk@pioneeredashks.org, or find us on LinkedIn. Thank you for listening to because you need to know, the reference podcast in knowledge management. My name is Soni Toneme. As an art administrator, Because You Need to Know has been my go to podcast and has helped me hone my management skills. Please consider sponsoring the podcast with your business.