Before the session, Vladimir Putin and Sberbank CEO and Chairman of the Management Board German Gref visited an exhibition on the revolutionary development of AI technologies.
The President toured the stands dedicated to the capabilities of and prospects for generative artificial intelligence. AI technologies are used for processing, generating and analysing various content, such as texts, images and music.
* * *
Excerpts from transcript of plenary session of the Artificial Intelligence Journey 2023 international conference
CEO and Chairman of the Management Board of Sberbank German Gref: Good afternoon, friends,
I would like to welcome all the guests of our discussion and the distinguished speakers of the AI Journey conference. This is the eighth edition of the conference, where guests from more than ten countries are physically present as well as online participants, more than 100 million people from 150 countries around the world. And every year, our conference becomes increasingly popular and the number of participants, the number of views as well as the number of countries included keeps growing no matter what.
Friends, today we are joined in this room by artificial intelligence specialists who are leaders in various fields related to the application and development of AI. Today we are going to talk – we have decided that there is cause for it this year – we are going to talk about large language models and generative artificial intelligence. Fortunately, the events of the past year have made it hugely important to put this issue at the centre of our discussion.
We would like to discuss what generative artificial intelligence based on large language models can give humanity, and how it will expand human capabilities and opportunities.
Also present in this room are our partners – companies developing artificial intelligence, members of AI Alliance Russia. I would like to thank all those present for their active work in preparing for this conference.
I would also like to express my gratitude to Mr Oreshkin, who has put a lot of effort into this, and to Mr Chernyshenko, to the Ministry of Digital Development, Communications and Mass Media, the Ministry of Economic Development and the entire Government for creating the right conditions and opportunities for the fairly rapid development of artificial intelligence technologies in our country. This is critical.
And I would like to thank you, Mr President, for taking the time to attend our conference as always. I would like to give you the floor now.
President of Russia Vladimir Putin: Mr Gref, friends,
I am delighted to welcome you to the international conference, Artificial Intelligence Journey 2023, everyone who is sitting in this room and watching our meeting online, namely scientists, engineers, programme developers, school and university students from Russia and more than 150 other countries. It is a huge audience.
Before delivering my address, I would like to express gratitude to the joint efforts of Sberbank, AI Alliance, the Russian Academy of Sciences, the relevant ministries, the heads of regions and, of course, members of the scientific and business communities.
It was mostly on this platform that we worked together to formulate a system-wide AI programme, and these efforts have produced the first results. For example, in the past few years Russia’s economic and social sectors have increased the use of AI solutions by 50 percent.
I will provide a concrete real-life example from the real economy sector, or more precisely, the fuel and energy sector. Gazprom Neft is using AI technologies to slash the cost of oil well development, to address complicated logistics safety issues on the Northern Sea Route, and other issues.
Investment in the high-tech AI businesses is growing in general. In this context, the Russian Direct Investment Fund is actively involving co-investors, putting its own money into domestic software companies, and helping them develop global cooperation and enter foreign markets, and today especially in Asia and the Middle East.
But an integrated assessment of efforts cannot be limited to business indicators alone. It also reflects the growing trust of people in a new, crosscutting, universal and essentially revolutionary technology. I have just talked to Mr Gref about this. Of course, with the introduction of AI in science, education, healthcare and all areas of our lives, humanity is opening a new chapter of its existence. I think this is perfectly obvious. It is a blessing that our people see how AI is making simpler and more convenient many routine processes, improving the quality of management and mechanisms of government services, and how it is being applied on a broader scale in organisations, enterprises and the work of regions.
I would like to cordially congratulate the winners of the AI Leaders prize. They include both young scientists and whole regions that, as I have already said, are proposing breakthrough solutions in AI and setting ambitious plans for progress. I cannot resist. I will simply name the winners. I know the media have already published their names but still. They are Alexander Korotin, Alexander Beznosikov and Valentin Khrulkov. The regions include Tatarstan, the Lipetsk Region, the Sakhalin Region and the Khanty-Mansi Autonomous Area – Yugra.
For the time being, few regions are developing all the opportunities of AI so energetically. I hope we will be more active in this area. When I say “we,” I am referring not only to the Government but also to the regions and industries, and individual plants.
Now I would like to tell you how the Government intends to work in this area going forward. I am going to tell you how the national leadership, the Government, the Executive Office, how all of us are going to develop domestic science, amend legislation and expand international cooperation, primarily in practical matters that are the main item on the agenda of the current conference.
I know that you are discussing a complicated but very promising topic – the future of so-called large language models and the further development of generative artificial intelligence. The young researchers and practitioners who work at Sber have just spoken about this. I can only add for non-specialists, such as myself, because I have had to grapple with some areas of artificial intelligence, that these systems make it possible to create images and texts. I will mention them once again. My colleagues have already described them in detail by citing examples.
These areas came into being literally a few years ago but began to dominate discussions and reveal their potential in just the year since our last meeting. Experts describe these ongoing processes as a revolution, a technological leap, a real breakthrough in the development of artificial intelligence. I have already said this is altogether a new chapter in the history of humanity.
Today, we have technologies with enormous calculating capacity and more advanced software architecture. They have absorbed all information, all knowledge that is available on the entire global internet.
Wide application of this technology has already become everyday reality. Frankly, not everything is ideal yet, but models are being continuously upgraded and carry out both applied and creative assignments. Generative artificial intelligence (as was just said) writes software code, generates component drawings and is gradually being introduced in the designing of structures and buildings. It reduces the time it takes to find the best molecules for medications and creates films, music and poetry. I would like to emphasise that these are just some of the areas where machines have matched or even exceeded human capabilities.
In effect, artificial intelligence algorithms or models are no longer primitive executors of our commands. A new generation of technologies are actually becoming our partner in the most diverse areas. This is a critical resource for businesses and states today to become super-efficient, which is what we need.
Using generative artificial intelligence, companies and enterprises are able to create products and services with special, unique characteristics to meet the demand of a specific customer. Naturally, such breakthrough solutions are opening up a path to creating a whole range of new business models. They use platform solutions to help reduce costs, introduce the principles of lean manufacturing, and significantly increase labour productivity.
In public administration, generative AI technologies make it possible to fully transition to data-based management, automate more administrative procedures, maximise the speed and reliability of decision-making processes, and thus radically improve and change the nature of many spheres that directly affect every citizen, primarily urban environment, mass transit, government service systems, environment, education, and healthcare.
I will address the social aspects of implementing groundbreaking solutions later, but I would like to highlight the following from the very start: artificial intelligence will not replace healthcare professionals or teachers, but it can act as a reliable and efficient assistant and give teachers more time to educate children, help doctors prevent and detect diseases early, and facilitate remote health monitoring.
Importantly, for citizens the everyday implementation of cutting-edge artificial intelligence is a modern human-centric social phenomenon offering a new quality of life and new opportunities for professional development. In fact, people can delegate to a machine many routine functions in daily life, production, management and even in creative industries, freeing them up to take on greater challenges. This means that modern jobs requiring substantial knowledge, skills, and competencies will replace low-skilled jobs, leading to higher wages for specialists.
In this context – I would like to point it out specifically – automation and widespread adoption of AI technology are not merely a resource for bolstering our scientific, educational, technological, and industrial potential, but also a crucial factor in rapid income growth, having a high-wage economy in Russia, and improving people’s quality of life as the primary objective of our advancement.
Earlier, in the first part of this event, our colleagues discussed developments in artificial intelligence and proposed asking a question – Mr Gref and I were caught off guard by the suggestion – to a machine. The most straightforward question that is pertinent to today’s agenda: how to make our country more efficient and make everyone happy. That is a question for artificial intelligence.
It is a broad question, but in order to make all of that happen in the near future, a presidential executive order will be signed and a revised version of the National Strategy for the Development of Artificial Intelligence will be approved as the initial steps towards achieving these goals. Taking into account the rapid growth of these technologies, a number of significant changes will be made to this document, and goals and objectives will be specified. This primarily concerns the expansion of fundamental and applied research in the field of generative AI and large language models.
I will present a number of specific solutions and proposals.
First, I urge the Government, the Alliance, and the Russian Academy of Sciences to devise a mechanism ensuring access for Russian scientists to existing and prospective supercomputers that are being built in the country. Graduate and undergraduate students, and schoolchildren involved in AI-related research and practical work should have special privileges when it comes to using computing infrastructure.
Second, I want the Government and Alliance companies to note that Russian supercomputers’ capabilities must be increased by at least an order of magnitude. This is crucial for the continued advancement of generative AI.
Three, it is imperative to significantly expand the training of highly skilled R&D engineers. The top-rated universities excelling in training AI specialists should have this task assigned to them. These include the Higher School of Economics, MIPT, ITMO University, and Moscow and St Petersburg State Universities, as well as Skoltech.
I propose that top-rated Alliance universities expand their master's and doctoral AI technology development programmes by September 1, 2024. Additionally, more students should be admitted to basic publicly funded AI programmes.
Fourth, we are open to considering the restructuring of research funding and allocating additional resources to research and development in generative AI and large language models provided our leading companies commit to co-finance these initiatives and to ensure the testing of relevant technologies, and, most importantly, delivering groundbreaking products that are competitive with the global benchmarks.
Notably, Russia is among the select few countries with its own generative AI technology and large language models, including Sber's products such as GigaChat and Kandinsky, and Yandex's YandexGPT and Shedevrum.
We must enhance our competitive advantage, and use this technology to create new markets and a whole galaxy of products and services. Doctors, teachers and builders, as I have already said, as well as agronomists, industrial and transport workers and government officials, to name a few, could be the first to use the most advanced technologies and enhance their efficiency.
In this context, I suggest using generative artificial intelligence for the development of large sectoral models and creating mechanisms for their practical introduction with a view to substantially increasing labour productivity and, hence, salaries in key branches of the national economy.
To speed up this process, I would like to ask the Government to draft, in cooperation with the Alliance, a special educational programme on the theory and practice of developing and using artificial intelligence with special emphasis on large language and generative models. This programme must be studied by the heads of major companies, federal and regional authorities, as well as universities and secondary vocational schools. It is necessary to introduce these instruments on the broadest scale. Importantly, education must start in the first quarter of 2024.
As you develop the five-year forecast of workforce requirements in cooperation with businesses, of course, I would like to ask you to very carefully analyse in what industries demand for existing specialties will change in the short-term perspective and where new professions, skills and competences will be required. Based on this analysis, the educational system must be given specific assignments to update the programme for the professional orientation and education of specialists.
The world is undergoing cardinal changes and we must stay ahead of them. No doubt, it is necessary to configure the entire system for supporting development and introducing large language models and solutions in artificial intelligence, and to create the infrastructure for their broad application. I would like to ask you to incorporate all these issues in the national project on creating a data economy, which is now being drafted.
Friends, the new generation of artificial intelligence is an outstanding achievement of the human mind. Nevertheless, it is essential for us to continue contemplating how humans will feel amidst machines in the broad sense of the word. Where do we set the limits on AI’s development? These ethical, moral, and social questions have sparked serious debate in our country and around the world. Some have even suggested putting on hold further work in the sphere of generative and especially strong artificial intelligence, which is expected to possess superpowerful cognitive abilities. As my colleagues said earlier, AI can smell, and has also shown early signs of empathy which is a distinctly human characteristic, it would seem.
Nevertheless, I am convinced that bans on technological development are not the way of the future, because it is simply impossible. Banning it is not an option, because AI will continue to develop no matter what. Even if we ban it, others will keep working on it, and we will fall behind. Everyone who has the capability to advance artificial intelligence thinks about it in this way.
However, it is crucial to ensure the safety and rational use of such technologies, and we should rely on traditional culture, among other things, because it is the most natural ethical regulator of technological progress, just like the ideals of goodness and human respect, as articulated by Tolstoy, Dostoevsky, and Chekhov, and outstanding science fiction writers like Belyaev and Yefremov. By the way, artificial intelligence can be asked to ponder ways to limit its scope of activity in order to avoid exceeding certain boundaries which could be harmful to humans.
The works of our outstanding writers have served as a moral compass for many generations of researchers, enabling our country to achieve scientific triumphs and to use these achievements for the benefit of people. This includes the peaceful use of nuclear energy, in which our country remains an undisputed leader.
I want to emphasise that in an era of technological revolution, cultural and spiritual heritage is a key factor in preserving national identity and, consequently, the diversity of our world and the stability of international relations.
But what facts are we already facing? This is now happening in practice. I think you know well that some Western search engines and generative models are often very selective and biased. They do not take into account and sometimes simply ignore and cancel Russian culture, for one. In simple terms, the machine is given some creative assignment and performs it using only the English language database that is convenient and beneficial for the developers of the system. Thus, the algorithm may tell the machine that Russia, our culture, science, music and literature simply do not exist. They are cancelled in the digital space, as it were. Later, they can do the same with other cultures and other civilisations, inserting themselves and emphasising their exceptionalism in this space as well. Artificial intelligence created according to some Western standards and patterns may emerge as this kind of a xenophobe.
I will repeat – many modern systems based on Western data are designed for the Western market. They fully reflect the aspects of Western ethics, norms of conduct and government policy to which we object. Of course, monopoly domination of such foreign innovations in Russia is unacceptable, dangerous and intolerable. Our domestic models of artificial intelligence must reflect the entire wealth and diversity of world culture, the heritage, knowledge and wisdom of all civilisations. This will make us only wealthier and more competitive. Naturally, our innovations should rest on our traditional values, the wealth and beauty of the Russian language and languages of other peoples in Russia.
We must certainly use Russian solutions in creating reliable, transparent and safe AI systems and involve specialists in the humanities in our common work. In this context, I would like to note a joint project of Russian Academy of Science institutions and Yandex. They have created an enormous collection of texts – the National Corpus of the Russian Language.
I also believe it is necessary for the authorities at all levels, libraries and archives to help create data sets for machine learning, and to provide them promptly and without charge. I would really appreciate it if my colleagues from the Government think of an adequate mechanism for this work.
We will continue to strive for Russia to become one of the world’s most conducive jurisdictions for the development of artificial intelligence and bold exploration of the technological solutions that everyone needs. Naturally, threats to the security of personal data must be minimised which we always insist upon.
To address this, at my direction, the Government has prepared amendments to the law on experimental legal regimes in the sphere of digital innovations. It will now be much simpler to establish such a mechanism. At least, I hope it will simpler and take half the time it took before. Many unnecessary bureaucratic requirements have been lifted. At the same time, though, liability for causing injury, death or property damage as a result of testing or using breakthrough technologies has been established.
I consider this approach to be absolutely balanced, and urge the Government to promptly submit the package of amendments to the State Duma. I urge the deputies to expedite the adoption of this document for the decisions to come into effect next year.
However, alongside legislative measures, voluntary consent of market participants regarding key rules of creating breakthrough technologies is of immense importance. When we discussed the importance of establishing the Alliance at our previous meetings, it turned out to be a very timely and valuable idea, and our colleagues are working effectively in this regard. I would like to highlight the special role of the National Code of Ethics in the sphere of artificial intelligence. It has been signed by domestic companies and organisations, as well as by participants from 16 countries worldwide. I believe complementing this document with a declaration on the responsible development of large language and generative models would be the right thing to do.
Russia’s experience can be instrumental in shaping international ethical standards in the field of artificial intelligence. We should strive for balanced and justified legal regulation that serves the interests of all, rather than select countries. I propose thoroughly discussing these issues during Russia's chairmanship in BRICS next year. We can certainly make arrangements for this discussion as the initiators of these efforts within BRICS. I am confident that our colleagues will support us. We must revisit this topic at the next Sber conference, which should be a full-fledged platform for regular discussions of the international AI agenda. We should address the ethical aspects of AI use at a dedicated international forum which will be regularly held in our country from now on.
To reiterate, Russia is ready to share its experience. I am aware that Sber, our research organisations, and universities are expanding collaboration with other countries. For instance, the typhoon diagnostic system, which was developed by the Far Eastern Federal University in collaboration with the Russian Academy of Sciences, has been translated into several languages.
Our plans include creating a repository for the code of the domestic platform and services that are indispensable for interaction of Russian and international software engineers.
I suggest considering granting our partners from friendly countries access to Russian large language and generative models. I am absolutely convinced that the technological world of the future must be multipolar, and we must build it together on the basis of trust and mutually beneficial cooperation. This is what we plan to do.
I have spoken about a great potential of this new generation of technology more than once today. The point of all our work and the main measure of technological progress is the ability to serve people.
Thank you very much for your attention.
Mr Gref promised that the participants in our meeting would talk about their projects in more detail. If there are any problems in organising this work, I will be pleased to listen to you. I see colleagues from the Government and ministers of the Russian Federation are present here. We will be glad to respond to your requests.
German Gref: Thank you very much, Mr President.
I would like to note that you mentioned in your remarks that AI should be used, among other things, to limit AI. You are absolutely right because according to the recent trend, the latest, more modern and advanced models have a function that makes it possible to set such assignments and fulfil them.
The problem of trusted AI is one of the big tasks. As you said, this happens when some data sets are missing, or when some transferred data are incomplete or have some exceptions. In this case, a more advanced model points to these drawbacks and shows how to resolve the problem.
The same applies to explaining and opening the AI black box to see how AI made these or other decisions. For the time being, people cannot get to the entire mechanism that is operating in the black box. In this case, later models are used to try to explain what principles of decision making underlie these or other conclusions.
Vladimir Putin: You know, this idea occurred to me just now during my speech. But I would like to express one more idea. It is obvious, there is nothing difficult in this. After all, humanity has drafted certain rules on the use of nuclear technology, including military use. It invented the rules for the non-proliferation of the carriers of this nuclear technology. Humanity managed to come up with these rules, so we can certainly find common solutions that would be acceptable for all and required by all.
German Gref: You are absolutely right. I think it is in the interest of all people, all humanity to display utmost responsibility in drafting these rules. People must fully control this process.
Vladimir Putin: Do you know when this will be possible? When all people sense the threat. This is not a joke. When they sense a threat in uncontrolled proliferation and uncontrolled work in this area, they will immediately want to come to terms. Of course, it would be best to do so without understanding what can be used in regard of other countries. If simply the awareness of common threats could evoke a desire to come to terms.
German Gref: Mr President, let me ask you a question.
First, what are your thoughts on the topics and aspirations that our speakers talked about here?
I would also like to ask you what dream you have.
Vladimir Putin: My dream is to see all that we discussed come to fruition.
One of the recent speakers said that we are currently witnessing revolutionary developments in digital technology, particularly in the realm of artificial intelligence.
I do not want now to provide specific definitions at this moment, but it is clear that revolutions involve the profound dismantling of old standards, rules and customs, leading to a transition, as they say today, to entirely new platforms and laws that govern the life of society and individual. Of course, what is needed here is not drastic changes of any kind but rather an evolution. However, this evolution must be organised quickly, with high quality and efficiency at all levels, spanning from businesses and municipal councils to the constituent entities of the Russian Federation and federal authorities, across all sectors of society. We must actively seek to achieve this, concentrating all our intellectual, administrative and financial resources on this if we want to secure a promising future for our country.
I want to wish you every success. Thank you very much.
German Gref: Mr President, colleagues,
First, I want to thank you for attending this conference every year. Your participation inspires a lot of people, including functionaries, government employees, members of the academic community and businesspeople.
I genuinely wish you success because the prosperity of our country depends, to a great extent, on your work.
As for your wish, there should be no doubt whatsoever that we have a good chance of seeing everything discussed as a dream here today materialise within a short timeframe.
Thank you very much.
Vladimir Putin: Thank you. Good luck and every success to you.