Home Latest Feeds Technology News Gergő Barta – There is MI, but it could help more

Gergő Barta – There is MI, but it could help more

0
Gergő Barta – There is MI, but it could help more

[ad_1]

One of the keys to the successful application of artificial intelligence is strong data management. There are several obstacles to the correct application of AI, such as the lack of a well-thought-out development strategy, but we can also mention the lack of specialists who understand AI.

Our cover interview was published in the February 14, 2024 issue of ComputerTrends magazine.

How do domestic companies use artificial intelligence? Answers to this question were sought in Deloitte Hungary’s survey, during which in-depth interviews were conducted with more than 200 decision-makers from 130 organizations. The research did not take into account ad hoc solutions used individually by each employee, only systems implemented or planned at the organizational level were dealt with. THE Hungarian MI Map 2023 most important results Gergő BartaDeloitte’s leading AI expert shared with ComputerTrends readers.

ComputerTrends: What are the main areas of use of AI in Hungary today?
Gergő Barta: The survey reveals that the three main areas of use of artificial intelligence in Hungary today are marketing and sales, management decision support, and the endowment of own new digital products and services with AI capabilities.

CT: In what other directions are organizations thinking about using artificial intelligence?
Gergő Barta: Here, too, we can mention the Top 3 areas. One is HR digitization, including, for example, the automated processing of resumes or the prediction of dropouts. This includes customer support and customer service. We can meet more and more chatbots in customer services that are already really connected with generative artificial intelligence. Thirdly, the directions under development include personalized customer experiences. The goal is to create as many personalized solutions as possible, mainly in the field of finance and online services, but also in the field of technology. I found the plans of a pharmaceutical company very interesting, according to which they see an opportunity to, in the future, based on the given laboratory results, prepare a medicine for the patients that, based specifically on the person’s data, helps to solve their problems.


CT: From what budget do organizations finance their investments in artificial intelligence?
Gergő Barta: The answer really surprised us, as 26 percent of the organizations said that they have dedicated AI resources. To tell the truth, we expected a smaller proportion. After that, we assumed that the AI ​​resources in this round are part of the technology or innovation budget. But no, the circle in question dedicates these resources to the development of artificial intelligence. Notably, 80 percent of organizations expect some increase in investment in AI technologies in the next financial year.

CT: Among companies with dedicated AI resources, what is the ratio of large companies and SMEs?
Gergő Barta: Most of them are large companies. SMEs are more characterized by the fact that the artificial intelligence budget is part of the IT, technological or innovation framework.

CT: What are the technologies where artificial intelligence is used the most?
Gergő Barta: The most popular is anomaly detection. At financial institutions, for example, artificial intelligence is used to investigate where fraud may occur in transactions. This includes the detection of IT security incidents, where the system may have been hacked or where viruses may have appeared, or quality assurance at manufacturing companies, the detection of defective products with AI or even machine vision. In summary, we can say that it is the detection of any outstanding points from data that may be an anomaly in the given system. There are many developments related to generative AI, solutions that can create sound, image and text. The more traditional use cases of natural language processing should also be mentioned, including either chatbots or the classification of text documents.

CT: According to those interviewed, what are the main success factors for the application of AI?
Gergő Barta: The main success factor is choosing the right use case. The answers revealed that many organizations have embarked on various projects due to the hype around artificial intelligence, which they then do not know what to do with. Not relevant, they caught a problem that could not be easily automated with AI, there was no suitable use case, business use. Many senior managers therefore made the mistake of starting to deal with the topic only because of the AI ​​hype. In addition to a good use case, the respondents consider the support of senior management to be important. The third success factor – by the way, one of the main conclusions of the research – is strong data management, because AI systems are fed by data. Adequate data management is the basic condition for proper storage, processing, protection, and integration of data. Many people mentioned the development of technical skills as well as the training and further training of the current workforce as a success factor.

CT: Do the senior managers who decide on an MI project have sufficient knowledge?
Gergő Barta: Acquiring new talents and knowledge is in sixth place among the success factors, so the problem is real. Although it was not revealed in the survey, we see from other sources that companies – while dealing with AI themselves – often use external help. Of course, there are organizations, such as large banks and large technology companies, which have their own AI excellence lab, so they have developed a competence center specializing in artificial intelligence in-house.

CT: Do organizations typically start MI projects based on a pre-created strategy?
Gergő Barta: Unfortunately, most do not. 68 percent of the respondents said that they did not have any strategy when they embarked on the MI project. This is probably why several interviewees complained that the project did not bring the expected satisfaction. The answers also revealed that few of them have a management framework that would cover the development steps well. What can be seen on the market is that AI developments mostly take place in an ad hoc manner, rather than in a controlled environment.

CT: How prepared are the organizations dealing with MI projects for risk management?
Gergő Barta: Based on the survey, I can say that they are moderately prepared. Respondents identified the biggest risks related to the ethics of AI systems. It is not clear when, in what way, to what extent artificial intelligence can be used, or how ethical the use of data is, how reliable the system is.

CT: Did the question of trust arise in the research?
Gergő Barta: Yes, we asked such questions. According to the respondents, trust is a key issue, transparency of use is important, i.e. knowing, for example, how and for what personal data is used. Responsibility and accountability are also cardinal issues from the point of view of trust. Those responsible should be named in case the AI ​​makes a mistake. The third key factor in creating trust is the explainability of decision support. This is important because most artificial intelligence systems are actually a black box, inside which complex mathematical operations take place, which the average person does not understand.

CT: What other obstacles and challenges do decision makers see in relation to the introduction of AI?
Gergő Barta: Security and data protection challenges, such as how data can be used under GDPR, were cited as the biggest issues. In addition, when it comes to IT, integration-complexity problems occur quite often, since many different technological ecosystems have to be coordinated.

CT: Overall, how do you evaluate the results of the survey?
Gergő Barta: It can be seen that organizations are spending more and more money on AI developments. It is widely agreed that artificial intelligence is essential for competitiveness. While most organizations jumped into AI development without a strategy, they realized along the way that this was a mistake. A good AI strategy helps to harmonize ideas, allows AI to align with business objectives, and this is very important. Many people see the possibility of reforming business models in generative AI and language models. In this field, there are already self-developed solutions on the market, not just boxed products. It can also be seen that the Hungarian market is not so risk-aware, but rather focuses on smooth operation, smooth business, and a return on investment as soon as possible. Organizations deal less with risks. It is also an important conclusion that there are few specialists dealing with AI, be it for example data scientists, data engineers, or engineers dealing with machine learning.

CT: Based on the study, how do you see where organizations should go when it comes to AI?
Gergő Barta: The introduction of artificial intelligence must begin with the creation of a clearly thought-out AI strategy focused on solving real business problems. It is very important to have data of the right quality. Developing data management capabilities, including data security and data protection, is key. If this is missing, there is nothing to build the AI ​​system on. We also see that, on average, 90 percent of an AI project can only be the cleaning of the data, bringing it into a suitable structure, data manipulation, i.e. the management of the data. I would also draw the attention of organizations to the importance of risk analysis. The lack of risk assessment may later lead to non-compliance, reputational risks, and may even be counterproductive.

[ad_2]

LEAVE A REPLY

Please enter your comment!
Please enter your name here