Artificial Intelligence — Beginning or End?

Intro

AI as a topic has been very present in our and global media space for some time now, and the intensity does not seem to be decreasing. On the one hand, we have predictions, like the one from Elon Musk, that artificial intelligence will become superior to human intelligence within five years, to the more conservative ones who warn that there is a kind of AI bubble deflation, given that many projects have not yielded the expected results and qualitative shifts have not yet occurred.

Some researchers suggest that AI could be a new — “general-purpose technology”, one of those technologies that will change the world, and it's functioning in almost all segments. Previous technologies that are considered to have had such an effect are the invention and mass use of the steam engine, followed by electricity, and lastly IT. If the predictions come true, significant social changes await us. Unlike the first technological revolutions, the last ones are happening faster and faster and it will be interesting to follow the possibility of adaptation, both of people on an individual level and of the social structures in which we live. In a more positive scenario, we can hope that AI will become, due to its high prevalence and low cost, something that everyone uses, as well as electricity, rather than a competitive differentiator.

Education and knowledge development

The challenges of developing artificial intelligence are numerous. A large number of engineers and technical experts are involved in the development itself, but the number of experts from other professions, ethicists, philosophers, etc. is also increasing. Acceptance of artificial intelligence in society is a complex psychological and sociological issue and has to do with public trust and information among other things. The last very interesting project comes from Finland (in collaboration with the University of Helsinki and the Reactor), and was recently presented in Croatia as well. Elements of AI is an education made just for the general population to answer the main questions: How could artificial intelligence affect your job or life? What is artificial intelligence really and how is it created? How will artificial intelligence evolve and affect us in the years to come? The general goal is to demystify artificial intelligence and all the prejudices and stereotypes that someone might have about it. It is nice to see that many Croatian public figures have been involved in the promotion of an educational project, the first goal of which is to educate 1% of citizens. As the president of the CroAI association, Mislav Malenica, said: “Understanding the world around you without understanding artificial intelligence will no longer be possible. Get involved on time, this is your chance. “

Numerous researches and data show how important it is to react as quickly as possible in the educational segment. A large number of professors leave colleges and go to the private sector. One of the consequences is after some period the creation of a smaller number of startups by students of these universities. The Dutch University of Eindhoven plans to invest 100 million euros, and thus “buy” 50 professors and focus its institute on the development of smart algorithms, robotics, and autonomous vehicles. Predictions about the first higher education institutions established with the specific goal of creating competencies in the field of artificial intelligence are also being realized. Thus, the first “AI Faculty” was established in Abu Dhabi — Mohammed bin Zayed University of AI (MBUZAI). Already for the first generation, they received applications from 2223 applicants from 97 countries. Among their employees are prominent experts from Oxford, MIT, and Google.

The current dominance of the United States is also confirmed by the fact that 59% of the top scientists in the AI ​​fieldwork right there. The fact that only 20% of experts come from the USA shows how good a job they have done in attracting experts.

China, with its investments in education and projects, has achieved exponential growth in recent years and has significantly increased the number of top experts. An important difference is that most of them still remain in the United States where they were educated.

Specificities of countries

The topic of artificial intelligence is very likely to have to be addressed by all, both at the individual, organizational and national levels. That is why the National AI Capital project was launched. As can be heard from experts at numerous AI conferences — “if we do not react immediately, we will not have a future.” The United Arab Emirates is the first country to establish a special ministry for the development of artificial intelligence, and countries like South Korea are developing national strategies, which include goals such as: developing 50 AI chips in the next 10 years. Taiwan has also used its expertise in semiconductor manufacturing to increase its geopolitical importance to its firms.

We mostly depend on private initiatives, of which the latest project from Poslovna Inteligencija should be highlighted — the establishment of a BIRD incubator intended, among other things, for the development of artificial intelligence ideas (applications are still possible).

New projects examples

While progress is not as great as some had hoped, news of progress coming almost on a daily basis cannot be ignored. Big companies have decided to take a serious bite out of the AI ​​cake, so Sony is combining AI with another new technology — drones (their new division is called Airpeak). In terms of building trust, IBM’s results are also encouraging, which speaks to a huge increase in satisfaction with chatbots. IBM, as one of the pioneers in the development of technology, will soon announce the first unmanned ocean-going ship. We could also see the rapid application of artificial intelligence in the current prediction of critical cases of patients with coronavirus. In one project, by analyzing the symptoms of over 4 million application users, new symptoms were identified before public health could react and make a diagnosis, without tests.

Bio-sciences are also experiencing a boom in successful AI projects, from genetics, chemistry to drug research. A sufficient indicator is an insight into the leap of scientific papers in the field of biology that has taken place in recent years.

A few days ago, a study was published in Nature in which Google’s Deep Mind took a huge step in determining the structure of amino acid-based proteins, a step that many scientists say will change everything in medicine.

On the other hand, we are rapidly approaching huge computational, economic, and environmental costs for incrementally declining advances in the predictions of some models. Thus, without new major scientific advances in some areas, it will take hundreds of billions of dollars to reduce current error levels by e.g. 10% to 1% error. Also, current methods use statistical methods to investigate the relationship between variables, which often requires that the conditions of ML system implementation be equal to those of the data studied. As a result, one of the main components of human learning has not yet been achieved — an understanding of cause and effect. Therefore, for example, diagnostic projects that rely only on the association of variables and through them the association of symptoms and diagnosis, are unable to separate the causes from the association. This leads to suboptimal or dangerous diagnoses sometimes.

Conclusion

All in all, it is not surprising that all predictions show a significant increase in investment in the development of artificial intelligence. According to some, the current, already quite high investments, will double in the next 4 years. There are growing voices and initiatives that warn of the importance of ethics in the development of artificial intelligence — the first case of wrongful arrests based on facial recognition software has been recorded. At the moment, 50% of the countries in the world allow the use of face recognition software and only 3 countries have a partial ban on their use. Clearview owns 3 billion photos supplied from various sources (for comparison, the FBI has a database of about 400 million images). It is to be expected that in the coming years there will be a great struggle on the legal-political field, given all the challenges that are coming. Fake news and deep fakes are certainly some of the first targets. Large investments are also made by the armies of individual countries, and the largest investments are received by startups that are in the defense segment of the industry. There is also the notion of AI nationalism, the political issue of allowing some startups to take over from other countries? China has set up a special fund of $ 29 billion just to reduce dependence on American technologies and reduce the current imbalance. At the beginning of 2019, the American president signed an “executive order” aimed at maintaining “American leadership” in the field of artificial intelligence. The impression remains that we are just at the beginning of various global political maneuvers and that things will rapidly intensify in the field of artificial intelligence.

For us, the only question remains, where will Croatia be at that moment and what topics will it deal with?

© 2021 Aco Momcilovic All Rights Reserved