Artificial intelligence: less informed internet users in Hungary

As a technology for collecting, processing, analysing and exploiting the vast amounts of data accumulated during digitisation, the widespread use of AI systems represents a step change in the already accelerating pace of digital transformation. According to most authoritative international and national research, AI systems significantly improve the operational efficiency and productivity of businesses. By replacing time-consuming, monotonous tasks, they also free up valuable human resources for higher value-added and/or necessarily creative human work.

For this reason, as part of a comprehensive survey on AI technology, we conducted a public opinion poll in Q2 2024 among Hungarian residents and businesses to map the attitudes and opinions of stakeholders on AI services and usage.

1. Public research

The survey found that, among online services, the most widely used AI systems were search (84%), social media platforms (77%) and video sharing (63%), but overall, over 50% of respondents use almost all of these services with varying degrees of intensity.

Respondents are also basically aware of the services that use AI solutions.

Virtual personal assistants, search services and video identification technologies were mentioned by more than 80%,

but solutions that support the online creation of text and images, translation solutions and social media platforms were also mentioned by more than two thirds.

Of the positive statements about AI made in the survey, considering the combined proportion of “fully agree” and “somewhat agree” responses, the statements “it makes learning and acquiring new knowledge more efficient”, “it can contribute to the development of health services and thus to the efficiency of health care activities”, “it saves time because it improves work efficiency” and “it frees up time by making administration, banking and shopping easier” found the highest levels of agreement among respondents, at over two-thirds.

Responses to negative statements about AI reflect the privacy and labour market concerns about the spread of AI in society. Of the statements with negative connotations, the statement “data that are not valid can lead to wrong decisions” stood out with a two-thirds agreement rate. More than half of the respondents agreed (either fully or somewhat) that AI “allows users’ data to be misused” and “eliminates jobs, jobs, increases unemployment”.

On the regulatory issues related to AI, respondents said that what is most needed are “regulations to protect personal data” and “regulations to limit the production of illegal content”.

2. Business research

35% of the businesses surveyed feel somewhat uninformed or fully uninformed about AI (rating of 4 or less on a scale of 10). A neutral position (rating 5-6) was taken by a third of respondents, while 32% of respondents said they were somewhat or fully prepared (rating of 7 or above).

The conscious and planned use of AI systems is only common to a small group of businesses.

The largest proportion of these businesses (8%) use AI systems for marketing and to support sales activities. AI support for other activities (RDI, administration, IT security, logistics, etc.) is not significant, and the responses suggest that enterprises do not plan to implement AI systems in these areas on a massive scale.

The main reason given by the businesses surveyed for not using AI systems was the low level/lack of perceived usefulness of such technologies (46%), while a third also mentioned a lack of knowledge about the opportunities available in this area (34%) and a lack of appropriate experts (33%). Security considerations were mentioned by more than a quarter of respondents (26%), while the cost of such solutions (23%) and the obsolescence of IT infrastructure (20%) were also mentioned by almost the same proportion.

Among the services using AI solutions, the most popular among responding businesses are internet search services (94%) and online financial solutions (87%). More than three quarters of respondents use online utility services (78%) and 61% also use online translation software. The majority are also present on social media platforms (56%).

Among the generative AI technologies, solutions that support the creation of text content are the most popular,

with a quarter of respondents saying they have used such a solution.

Based on the responses to the question on the impact of AI on business operations, it is clear that a low proportion of respondents think that AI solutions would have a somewhat or significantly negative impact on their business, while only less than a third were somewhat or very positive about the impact of these technologies on their own business.

Strategic awareness of AI is low among Hungarian enterprises, as the vast majority of respondents do not have a digitalisation strategy, a data strategy or an Industry 4.0 strategy, and although—contrary to our previous research—the majority of respondents said they have a corporate strategy, it does not include digitalisation, and even less so corporate plans and objectives related to AI.

An important consequence (or even cause) of the low level of integration of AI solutions in enterprises is that 8 out of 10 enterprises do not have a specialist who is partly or entirely responsible for the deployment of such solutions. In addition, 90% of the enterprises surveyed had neither launched internal training nor supported employee participation in external training to acquire knowledge about AI.

Two thirds of Hungarian enterprises agreed either somewhat or fully that “the use of AI can lead to wrong decisions if the data used for data-driven decisions is not controlled”, “the development of digital and other cognitive skills needed to use AI is essential” and “the use of AI saves time by improving work efficiency”. The majority also believe that AI can create new markets, new industries and reduce the scope for error in the way people work.

Less than half of the respondents said that AI would eliminate jobs and job opportunities due to automation.