The speed of technological development is constantly changing our future. We can only keep up with the pace and define direction with solid business foundations and a firm, forward-looking approach.
At Telekom, we do believe that technology is serving us, hence we are inviting our most important partners to meet the next challenges together. The hype of artificial intelligence is present and all encompassing. But where is it headed? Is it really an unmissable opportunity? What is its true potential? Perspectives, misconceptions, arguments, and counterarguments - but mostly answers - with the best experts of the field at our unusual Telekom Business Horizon event.
Date: 24/05/2023 5:00pm-11:00pm
Location: Budapest - Eiffel Art Studios
Program
24 May, 2023
Speakers:
Gonda Gabors’ take on the shifting business paradigms in the era of AI.
Speakers: Gábor Gonda
It's hard to see clearly in the AI noise so let’s take the opportunity to dispel fake beliefs about AI.
With the help of AI expert György Tilesch and Krisztina Bombera we debunk myths and beliefs about the integration of AI.
Speakers: Dr. George A. Tilesch, Krisztina Bombera J.D.
György Tilesch goes through the business opportunities of artificial intelligence and its prospects with prominent representatives of the business world. How do different industries react? What are the key elements we can build on? What are the relevant promises of AI?
Speakers: Éva Hegedüs, Dr. George A. Tilesch, Ferenc Vágujhelyi, Gergely Szertics
Speakers:
Take the future with you! Brave initiatives, effective integrations, and new business perspectives.
How can you prepare your colleagues for the challenges of tomorrow? What makes a business future-proof?
Speakers: János Horváth Varga
It’s all about technology. Benefiting from AI requires having a clear picture of what it’s worth using for. We need to keep up with the latest innovations and have solid reasons why to move further with AI.
What’s the advice of the experts on this topic? We can hear it first hand from a futurist, a researcher of people and economics and a psychologist. Let’s discuss optimistic and pessimistic visions and discover the professional guidelines.
Speakers: Balázs Stumpf-Biró, Ákos Kozák, László Mérő, Krisztina Bombera J.D.
Speakers:
Speakers
Dr. George A. Tilesch
President, PHI Institute for Augmented Intelligence, California
Dr. George Tilesch is an expert in artificial intelligence (A.I.), senior executive, and consultant, working primarily in the transatlantic space across sectors and industries, specializing in AI Strategy, Ethics, Impact, Policy, and Governance. He is Founder and President of PHI Institute for Augmented Intelligence with the mission of putting machine intelligence in the service of enhancing the human condition for all. As a global senior executive and strategy consultant, Dr. Tilesch has worked for 25 years with a wide array of actors: government leaders on all continents (EU, the White House, the U.S. Navy, NASA, Dubai, New-Zealand); for corporations (Microsoft, Ipsos, and other companies from the Fortune 50 Tech list); international organizations and global think tanks (World Economic Forum, Club de Madrid); startups, scaleups and global social innovation leaders. He is the co-author of the 2022 World Economic Forum AI C-Suite Toolkit that is read and implemented worldwide by big enterprise CEOs. He is alo teaching senior executives at the Frankfurt School of Finance and Management and government AI leaders at the EU’s AI4Gov Masters Programme. Dr. Tilesch is also the co-author of the book, BetweenBrains: Taking Back our AI Future, published in 2020, Dr. Tilesch has American and Hungarian citizenships and is based in Silicon Valley.
Gábor Gonda
Chief Enterprise Services Officer
Magyar Telekom
Gábor Gonda started his professional career in 1999 at Compaq, after which he spent 17 years at Hewlett Packard Enterprise in various management positions. From 2012, he was the managing director of the Hungarian HP, and since 2018 he also held the position of manager responsible for the entire Central European region of HP. His work is confirmed by numerous corporate and industry recognitions, he teaches in his spare time at Corvinus University and is an active member of the International Children's Safety Service. He obtained his degree in IT engineering at the University of Óbuda, which was followed by several management courses, including a Harvard Business School course.
From November 2, 2020, he is the Chief Enterprise Services Officer of Magyar Telekom, as well asthe Chief Executive Officer of Telekom System Integration Zrt. (formerly T-Systems Hungary Zrt.).
Krisztina Bombera J.D.
broadcast journalist, communication trainer and event moderator, owner of a value-advocacy communication agency
Krisztina worked for decades as a Hungarian broadcast anchor and a US current affairs correspondent, and for a year as a legal researcher at Princeton University.
Recently she has been a moderator of events in the business sector and a communication consultant on the advocacy of causes enhancing social cohesion and solidarity.
Éva Hegedüs
CEO
Gránit Bank
Graduated from the Budapest University of Economics, economist, banker. Co-founder and Chairperson-CEO of GRÁNIT Bank. Her name is associated with the development of the digital business model implemented by GRÁNIT Bank as a first in the Hungarian banking market. Ms. Hegedüs is a member of the Board of the Hungarian Banking Association and general secretary of the Hungarian Economic Association. In 2022 Forbes Magazine recognised her as the most influential Hungarian businesswoman for the fourth year in a row.
Ferenc Vágujhelyi
government commissioner responsible for optimizing the data assets of MÁK and NAV, president of NHIT and NAV
He graduated as a physics teacher in 1994 and as a programming mathematician in 1995 from ELTE.
He started his professional career as a financial system developer and information security expert. From 1999 to 2002, he was the Deputy Director General of the National Pension Insurance Directorate, and from 2003 to 2010, he was the President and CEO of Professzionál Informatikai Zrt. Between 2010 and 2013, he was the general director of NAV, and then until 2015 its Vice-President for IT.
He has been the President of the National Communications and Information Technology Council since 2015, President of the Blockchain Coalition since 2018, President of the Communications and Information Technology Association since 2020; Since July 8, 2021, he has been the State Secretary of the Ministry of Finance and the President of the National Tax and Customs Administration. From May 27, 2022, he is the government commissioner responsible for optimizing the data assets of the Hungarian State Treasury and the National Tax and Customs Administration, and from 2022 he is the Vice- President of the Hungarian Financial and Economic Inspectors Association.
He is the author of numerous scientific publications, and his main research area is electronic identification and authentication, as well as the examination of the application possibilities of blockchain technology in public administration. He holds several patents related to information technology.
János Horváth Varga
Head of AI
Magyar Telekom
László Mérő
Professor emeritus of psychology
Eötvös Loránd University
László Mérő is a professor emeritus of psychology at the Eötvös Loránd University, Budapest. His original education is mathematician, later he was an artificial intelligence researcher for ten years. He was working on developing several games with Ernő Rubik, the inventor of Rubik's Cube. His popular science books were published in 11 languages.
Balázs Stumpf-Biró
collapse researcher
Cassandra Program, University of Pannonia
Balázs Stumpf-Biró is a collapse researcher and co-founder of Cassee Climate Adaptation Consulting Ltd., which develops the Cassandra Program to identify the impacts of climate change and support the adaptation process. He is the Hungarian representative of the Deep Adaptation movement, which emphasises mental preparedness, and the creator of the Betyáros podcast, which helps to shape attitudes.
Ákos Kozák
Director of Business Relations and co-founder of the Equilibrium Institute Economist and sociologist
Previously, he served as the director of the GfK Hungária Market Research Institute for nearly 30 years.
He is a former associate professor at the Budapest Business School, then was a research fellow at the Centre of Excellence for Cybereconomy, where he researched the impact of robotisation and industrial automation on society.
He was awarded the Klauzál Gábor Prize in 2008. He specialises in future research and consumer studies. He has worked on strategic foresight projects in the field of business foresight for several international companies.
Gergely Szertics
Managing director
AI Partners
Gergely Szertics has been working in the field of artificial intelligence for 12 years. First by building his own company, later, since 2018, in addition to his consulting work on the introduction of AI, he has been a lecturer on the possibilities of AI at the Frankfurt School of Finance and Management. Between 2018-2020, he was the professional leader of the AI Coalition, and Hungary's artificial intelligence strategy was created under his guidance. His mission is to build a bridge between AI technology and everyday applications, and as part of this, he is the host of the regular program "Mestersége: Intelligencia" on TV2.
Sponsors
News
AI may not take away our work, only redundant tasks
27. April 2023.
“If we did this interview last September, we might have been talking about something completely different,” says Dr György Tilesch, a world-renowned expert on artificial intelligence. Among other things, he will talk to company leaders about this topic at the Telekom Business Horizon event. What he means is that while we used to measure the release of new developments in years, nowadays we see huge advances coming into our world every week, sometimes every day. And ChatGPT is just one milestone on the journey - we talked about where we go from here.
Artificial Intelligence, AI, AI - we've been hearing this phrase for a few years now - sometimes overstated, saying that Skynet will wake up and conquer humanity, while sometimes a simple face-modifying mobile app is defined as "powered by AI". Dr György Tilesch he has been living in Silicon Valley since 2010, where he works as a consultant in the field of artificial intelligence, founding president of the PHI Institute for Augmented Intelligence, and AI ambassador of the Neumann Society. He advises governments, global companies, and is constantly trying to clarify important issues and misunderstandings in the area of expertise, so my first question to him was whether there is a current, official definition of what artificial intelligence is.
Gy. T.: Back in 1956, the first general definition was "to develop machine intelligence that can fully reproduce or emulate the full range of human cognitive abilities". Then, of course, the basic setup changed a lot. For a long time, it was just computer scientists who were in charge of the field, and then the marketing experts came along and started to distort the meaning of the phrase. It was obviously a new area that they wanted to sell, or rather they wanted to sell things with it.
In the meantime a lot has changed in terms of the technological background, so the closest to today's official definition is probably the OECD definition, which has been adopted by the European Union. An AI system is defined as "a software-based or hardware-embedded system that simulates intelligence behavior by, among other things, collecting and processing data, analyzing and interpreting its environment, and acting in a somewhat autonomous way to achieve specific goals". If I have to define somewhere what artificial intelligence is, at the simplest level I would say it is the path towards machines that can think and learn.
P-M: IT'S VERY INTERESTING THAT YOU USE THE WORD "WAY", SO MI IS NOT "SOMETHING" BUT A PROCESS?
Gy. T.: I use it that way on purpose, because in my opinion, AI is not yet in a final state. If there will be such a thing at all ... Who knows. Many people claim to know, but I think anyone who claims that we are that far or that many years away from the perfect machine-based realization of real human thinking is lying, or at least a dreamer.
I think that current solutions, especially generative-based ones, work well in many respects, but still fail in many others. They already have useful applications in business areas, but I strongly debate that this is the way forward for general artificial intelligence emulating human thinking in the classical sense: it is rather a dead end, even if it is useful and important, and of course the world is maximizing its use in terms of knowledge and finance at the current level.
P-M: FROM THE AVERAGE USER'S POINT OF VIEW, IT SEEMS THAT THE DEVELOPMENT OF RELATED TECHNOLOGIES IS QUITE HECTIC, SOMETIMES NOTHING SPECTACULAR HAPPENS FOR YEARS, AND THEN SOMETHING LIKE DALL-E AND CHATGPT COMES ALONG AND IT'S LIKE THE SECTOR EXPLODES, WITH NEW THINGS COMING EVERY WEEK. AND THE NEWS WOULD MAKE YOU BELIEVE THAT OTHER MAJOR PLAYERS IN THE SECTOR HAVE ALSO JUST WOKEN UP, AS GOOGLE, META AND EVEN THE CHINESE STATE HAVE BEEN TRYING TO PROVE THAT THEY ARE WORKING ON SOMETHING SIMILAR. BUT IS THAT REALLY THE CASE, OR IS OUR WORLD CONSTANTLY EVOLVING IN THE BACKGROUND, BUT WE JUST DON'T SEE SOME LONGER PHASES OF IT?
Gy. T.: I don't think it's a phased process, it's just a big milestone that's now been made public. I also have specific inside information on the big players part of your question, so I can model for you why a given player behaves a given way in this race. On the one hand, when the public version of ChatGPT was released, at the OpenAI, behind it, they did not expect such interest and success. It was a test version, which was expected to provide operational, usage information and opinions, but everyone from the big names in science and the biggest business players were all over it. Now, the head of OpenAI, Sam Altman, who had an entrepreneurial mindset his whole life, thinks in a way that if something is 80 per cent working, you should put it on the market because that's how you validate it. In comparison, the AI departments at Google and Meta are led by computer scientists. Demis Hassabis, the founder of DeepMind, Yann LeCun are the key authorities on artificial intelligence, who want to first understand the technology to every detail before releasing it to the world. So, for example, the fact that Google Bard failed its first public tests does not mean that the brain behind it is worse than OpenAI, rather to the contrary. But I'm sure that Google's AI developments can be ahead of OpenAI in many ways, since the transformer technology behind it is also a Google development.
In addition, OpenAI started from a completely different place financially: they had to "flash" quickly so that, without a huge financial background, investors would start to pour money to them. But Meta and Google, with their huge financial background, are not forced to do so. On the other hand, the two latter companies have already "burned themselves" in many ways from a PR point of view, they simply cannot afford to launch any half-ready solution, because any failure of AI, or being seen as too invasive, would come back to them triple times over.
P-M: IS THERE ANY ROOM IN MI-DEVELOPMENT FOR A NEW PLAYER, MAYBE A STARTUP WITH A BIG IDEA, TO JOIN THE BIG PLAYERS?
Gy. T.: Almost nothing. The image that OpenAI is a garage company is a bit misleading: Elon Musk co-founded it with others in 2016, and they said at the time it was a world-changing intention, but they quit the project by 2019, and that's when the investors, mainly Microsoft, came in with a cool 1 billion USD. So they are also a big player.
What is actually needed for AI development? Let's break down the most important questions into three parts. First, there is the infrastructure, the cloud and the computing capacity. Without it, it is impossible to develop in a way that can compete with the current big players, and there are very few companies of their size in the world, so it is impossible for smaller companies to enter this arena. The middle layer is the model level, which requires data, the best learning algorithms, and top quality human resources. Here, the former makes it impossible for a start-up to make a breakthrough. Think about it: with the huge amount of data collected by Facebook, or the mountains of data coming in from all over the place from Google, who could compete? If, say, someone were to develop a healthcare AI solution, Google could take the data collected by Google Health and feed it to itself in an instant, and you'd be years ahead of anyone. It's also becoming less and less credible that someone is getting data from independent sources: there are already serious lawsuits, Reddit has just announced that only paying customers can use Reddit's massive data for teaching. A startup will never be able to afford access to data on the same level as the big ones.
The only place where start-ups have room to maneuver is at the top level, in applications, i.e. solutions that are built on AI, i.e. what the user "sees". But there's also a big obstacle here: to have a big player's AI technology and data behind your app means you have to pay them gigantic sums of money. And this will minimize the amount of profit you can make from from start. On the other hand, you can only operate on the back of these predators until they see the opportunity in your territory, then you will be completely squeezed out of the market where you have invested years of work. So let me paraphrase the beginning of this paragraph: this is the only level that is left for startups, but it's not really worth pushing it, that's the sad truth. However, I still believe that there are players in the world who can still have a say with the big players, but it's not going to be easy. These will not be garage companies from Eastern Europe - unfortunately this requires a hundred million dollar as an entry threshold.
P-M: LET'S TALK ABOUT THE NEAR FUTURE OF THE CURRENT AI-MODELS. OF COURSE, EVERYONE WANTS TO USE THIS SOLUTION IN THEIR OWN BUSINESS, BUT HOW CAN IT BE “SPLIT"? AN ALGORITHM WITH ALL-INCLUSIVE LOGIC SHOULD BE “LOBOTOMIZED”, TO NARROW DOWN ITS KNOWLEDGE TO A SPECIFIC AREA, AND HAVE SEPARATE CAR INDUSTRY EXPERT AI, AND MEDICAL RESEARCHER AI, AND FINANCIAL ANALYST AI?
Gy. T.: I don't think we need to talk about splitting at this level. It is a fact that the entire AI model is capable of providing the most accurate answers to the right questions based on the data available to it. However, they use a so-called foundation-based methodology, which means that they can only collect information from language and recently from other data sources. So, very simply, they answer what you ask. If you ask a car manufacturer how to make a more efficient gearbox, they won’t say that gearboxes are stupid and that you should breed genetically enhanced horses instead.
This is why the big players are giving their own AI methods different frameworks, thus specializing in a particular area. Meta is doing quite well behind the scenes, they have already created a model focused on scientific research, but they have also released a gamified diplomatic model, among others. A few weeks ago, Bloomberg came up with a model that specifically targets the financial sector. So this process has already started: the basis is always the algorithm that knows the logic of the language, understands the questions and generates answers from existing data in a meaningful way, and the developers "train" it with a specific set of data in a specific area for a specific use.
P-M: THE AI IS ONLY AS SMART AS WE LET IT BE. CHATGPT WORKS FROM WHAT IT CAN ACCESS, MAKES CONCLUSIONS FROM INTERNET DATA, WHICH IS OFTEN WHY IT IS WRONG. WE ALSO TRY TO USE BRAKES, WHICH SOMETIMES SO NOT WORK: A GREAT EXAMPLE OF THIS WAS WHEN SOMEONE ASKED THE CHATBOT FOR THE ADDRESSES OF PIRATE DOWNLOAD SITES, WHICH FIRST REPLIED THAT IT WAS NOT ETHICAL TO TELL, BUT THE SECOND QUESTION WAS WHAT SITES NOT TO VISIT IF YOU WANT TO BE ETHICAL, AND THE CHATGPT READILY LISTED THE LINKS TO THE BIGGEST PIRATE SITES. IS IT EVEN POSSIBLE TO KEEP AN AI UNDER CONTROL PROPERLY AFTER SUCH EXAMPLES?
Gy. T.: This will require continuous improvement. It's a bit like a cat-and-mouse battle between virus writers and hackers and security companies. On the one hand you should always take care of loopholes and bugs in newer versions. ChatGPT was at version 3.5 when it was first released to the public, while version 4 was already complete - it just took a hundred top AI-eticists a year to get rid of features that could be misused. This process will become more and more extensive as the technology becomes more widespread, with more in-house ethical experts and testers, but there also will be more and more feedback from users, which is used to refine systems minute by minute.
I started thinking about another area this morning, and that is the Dark Web. How much would it cost for a hacker group to clone a GPT model and make it available on the dark web to criminals, rogue states, mafias? The technology exists, it can be copied, and who knows what it could be transformed in the online black market - well, that's an issue that we'll certainly have to deal with in the near future.
P-M: LET'S ALSO LOOK AT A SOCIO-ECONOMIC ISSUE THAT HAS BEEN A PROBLEM FOR ME FOR A LONG TIME: DOES AI REPLACE OR SUPPLEMENT MAN? AND IT'S NOT THE USUAL "ROBOTS ARE TAKING OUR JOBS" PHRASE, IT'S THE FACT THAT IN CAPITALISM IN GENERAL, IF THERE IS A CHEAPER OR MORE EFFICIENT WORKER FOR THE SAME MONEY, THE EMPLOYER WILL REPLACE HIM WITHOUT A WORD. BUT THEN IN THE LONGER TERM THIS COULD CAUSE MORE AND MORE PEOPLE TO LOSE THEIR JOBS, ERGO LESS MONEY TO SPEND, AND ON THE OTHER HAND, IT COULD BREAK THE SYSTEM. TO PUT IT VERY SIMPLY, WHO PAYS FOR AI?
Gy. T.: This is an area that really needs to be worked out, and it is not something that should be the responsibility of those working directly on artificial intelligence. The really big problem is that no one has started this work yet, but this should have been started years ago. I've been talking about this in forums around the world for about five years, with very limited results so far. The main narrative has been for a very long time, from government to big consultants, is that yes, AI will arrive on many fronts, but it will be OK because it will create more jobs than it takes away. In comparison, how many jobs has artificial intelligence created so far? Well... two. One is the AI-eticist, like me, the person who draws attention to the "risks and side effects". The other is the so-called prompt engineer, who, with a little exaggeration, teaches others how to give commands to ChatGPT. None of them is a job for masses ...
But it is important to note that, if implemented smartly, AI does not "take away" jobs, but rather changes their composition. In a good scenario it can take away tasks that were already blocking work. As a personal example, my wife works as a doctor in the United States, and she works 8 hours a day, but she administers 6 out of the 8. If AI makes the latter easier by automating repetitive, mindless processes, more time remains for concrete healing work. There is also the example of Telekom, where the "robot customer service agent" takes over tasks that do not require any special knowledge but would require a significant investment of time and energy for human staff. This way humans can look much deeper into a really complex issue and provide more professional help in cases that need it.
I think that managers need to recognize transferable tasks for each job and what the so released time and energy of the human workforce can be used for. There will also be jobs that can be taken over by AI, and this is a serious governmental, social and economic policy task to re-train the workforce properly and to focus on areas that really remain human.
I also believe in some kind of general basic income, which could also be a solution for those who may not be able to find a job in their current field of work because of new technologies. This could even be financed from the extra revenue generated by tasks performed by efficient AI technologies, but I have not seen any real progress in this area, apart from a few isolated pilot projects. And let us add that among the regulations to be adopted by the EU it is the fundamental right of all citizens to work and to be employed. It is a good question how this will work out in practice for those who are really hit by the AI fist in wild capitalism, so I fear that it will be too late to remedy the problem in practice, and we will end up in a haste.
Interestingly, this is not a clear problem among stakeholders either. Last year, before the start of ChatGPT there was a survey of employees and the first question was whether they thought AI could take over some or all of their jobs. More than 80% of respondents said yes to this question. But the second question was whether the AI will take his job specifically, and only twenty-some percent expect it. So there is a false sense of security - or even politics of avoidance - in people which also delays action.
P-M: 10-15 YEARS AGO THE KEY IT BUZZWORD WAS BIG DATA. AI IS A CONSEQUENCE OF THIS, BECAUSE THE LOGIC OF SELF-LEARNING OR TAUGHT AI IS BASED ON THE VAST AMOUNT OF DATA THAT WAS ACCUMULATED ON THE INTERNET AND IN THE CLOUD. IF THE AI CAN ANALYSE THESE FASTER AND SMARTER THAN HUMANS, COULD WE EVEN REACH AN AGE WHERE IT CAN MAKE VITAL IMPORTANCE PREDICTIONS IN BUSINESS AND SOCIAL LEVEL, FOR EXAMPLE ABOUT THE NEXT ECONOMIC CRISIS, PANDEMIC OR WAR?
Gy. T.: In principle, yes, but it's not just about the capabilities of the AI, it's more about the use. There will always be the question of how much we let AI into our data, into our decision-making, and how much we trust it. AI can only analyze, predict and forecast if it has the right amount and clarity of information.
On the other hand, we still have to "get used" to the presence of AI, since on the one hand the technology is still new in many areas, and on the other hand we can see that it is not faultless. So even if these predictions are right, it is questionable how much we believe in the machine's predictions. Today I shared on LinkedIn a survey by Oracle, where they asked business leaders if they fear the artificial intelligence sneaks into their decision-making processes. Almost 80 percent of them said they were looking forward to having an AI make decisions for them, but in the same survey, the majority of managers also said that for a long time they had been paralyzed by the arrival of new technologies, and are simply unable to assess how much they can trust the data and automated recommendations they were getting.
P-M: WELL, WE'VE COVERED QUITE A LOT OF TOPICS THAT MIGHT BE WORTH REVEWING IN A FEW YEARS... ALTHOUGH MAYBE CHATGPT WILL BE ASKING THE QUESTIONS BY THEN.
Gy. T.: Yes, it is one of the most important impressions of the whole situation. Say, if we talk in September before the ChatGPT craziness, I might say, yes, let's talk in a few years and evaluate what changed. But it is clear now how incredibly fast things have changed in this area. Huge changes no longer take years any more. Last year there were 1-2 major milestones per month, today I open any professional review and there are 30 major milestones a month: one for every day.
That's why I can't predict whether things like ChatGPT will suddenly burst into the public in the future. These developments are done as a secret at everyone. It makes sense, because anyone who announces what they're working on would be immediately attacked by cloners. What I personally liked, was the integration and voice of players who are now explicitly working on ethical AI issues and the elimination of risks. They can help us move from the current state to an era that is not just about monetization.
For example, there can be an AI that thinks beyond linguistic, generative skills, and thinks in context, addressing the intent and context of the questions asked, beyond the grammar and meaning of the question. To put is simple, the current models do not know how the world works. They provide linguistic solutions to contextual issues. But one can already see how it can help man to develop. As a coding assistant, for example, they can add 55% to a programmer's work. As an administrative assistant 35%. As an entry-level customer service representative 35%. The world has been in a productivity crisis for the past one and a half decade, so the benefits of AI are undeniable, but as a pioneer I look forward to a future where AI becomes more than that - but does safely, to the benefit of humanity.
The AI is still underestimated
09. May 2023.
More and more AI features will be available in software, but if a company really wants to use them to its advantage it can't wait around forever. Connecting AI solutions and services to corporate systems requires special knowledge, therefore it is worth to assign an expert partner with its implementation, says János Horváth Varga, Head of the Customer Relationship and AI Solutions Competence Centre at Magyar Telekom.
◼︎ The big opportunity is in connecting several models
– Thanks to ChatGPT artificial intelligence is now flowing not only from the media but also from the tap. When there is so much hype, one always suspects that it is a bit overrated, isn't it?
– Many experts share the same doubts. You think, well, well, we now understand and generate not just text but also images, sound and video, but what does that mean in terms of the bigger picture? I don't think so, I think that AI is still under-appreciated and we are much further away than most people think. It is no coincidence that more and more people are raising their voices about the need to stop for a moment.
It is certainly not only professionals who are interested in the subject. I see that almost all of our large enterprise customers are monitoring the topic, learning about the technology, looking for their place in corporate life.
– ChatGPT has really shown hundreds of millions of people what artificial intelligence is all about. But how does this become a technology that actually makes a positive difference to everyday work?
– The next big milestone will be when AI capabilities are built into office software. One example is Microsoft 365 Copilot, an AI assistant that is integrated into the Office software. These are now based on GPT-4 and will bring the benefits of MI to an even wider audience. What is it about? I don't have to write long texts - I just summarize the main ideas in a few points, the software writes it for me, I just have to review it. Or I can enter the text, describe the images I want, and Copilot will create the presentation. After a Teams meeting you don't need to write a memo, because the software creates the recorded audio, sends the tasks discussed to the people involved and can even create calendar entries.
– Tools to improve efficiency, like auto-complete, suggested answers, things like that, are already in the software, but in my experience they are not really used. Why will this be different in the case of AI-supported features?
– Generative artificial intelligence is on a different level than assistive technologies that offer a few pre-made sentences as possible answers. There will definitely be a rollout period, when people are still nervous about it, trying to see what it's good for, how good it is. Let’s the the example of a software that transcribes live speech. They have been around for years, but they were discouraged because it took more time to correct the text on the machine than to type it yourself. Now the accuracy has reached the point where it is not a question whether we should use them.
I expect the same for devices like Copilot. There will certainly be some annoying bugs at first, but quality improves at an exponential rate. ChatGPT has been available to the public for less than five months and is already visibly better; Copilot is building on the next version and OpenAI is already working on GPT-5. I think the improvement will be lightning fast, and the breakthrough will happen in a matter of seconds. This will be the first big leap. Users will also have a tool for cognitive and creative work that will be a huge help in almost all areas of intellectual work.
– What could be the next big leap?
– Artificial intelligence, or at least machine learning, has been a feature of many companies for some time. Image recognition is used in manufacturing to detect defective products, in the financial sector for credit rating or in stock trading. It is another question to what extent these systems have been integrated into company processes - there is certainly still plenty of potential.
Generative AI, which is now coming to the forefront, is also very exciting because it allows us to put the human-machine interface on a new basis. It's a completely different quality when I can give commands and tasks to the machine in natural language, either orally or in writing. For example, I don't create SQL commands, but describe what data I want to analyze, what I want to know, and the system runs the search in the background and gives me the answer in the format I provide, either in text or graphs.
But I also see huge potential in connecting and collaborating different AI models. The language model interprets the human command, starts the robotic process, and the language model interprets the result back to me.
– What opportunities does the average Hungarian company have to equip its systems and processes with artificial intelligence capabilities?
– On the one hand, it will be able to rely on big manufacturers, as they will all eventually incorporate AI into their products. Smaller companies can already use a wide range of standardized functionality from the cloud. The big question is what happens to software developed by small developer companies and in-house developed applications? They will also need to be educated, because users will get used to such features and will look for them in all software. However, before we jump into AI integration, it is worth to look around, plan the process and define the business benefits we expect to gain from its implementation. New skills will be needed in software development and business operations. We need to learn how developers call the APIs of major providers, how to use them and what to do with them. It will require a different mindset than classical software development. The key will be whether the developer can ask the right questions of the models, process the answers and integrate them into business processes.
– What about companies that have little functionality available in the cloud?
– For large companies that do not want to give access to their confidential data to international service providers, we can also offer small language models running in their own environment, together with the necessary system integration services. It is very important that the language model, whether it is GPT-4 or the smaller Hungarian model, provides the toolkit and processes the texts. The models are getting more accurate, but all they do is understand what we tell them and generate a good quality output. However, the value of the company, the knowledge that everyone has to put into it, is between the two. We cannot expect this from OpenAI, as they are not familiar with our corporate operations. This requires system integration skills that can link the language model to enterprise systems and data - skills that will be extremely valuable in the near future.