Are you sure?

AI Community Needs Coordination to Thrive

There’s a lot more going on in the realm of artificial intelligence in Hungary than you might think. However, isolated efforts must be coordinated so that the country can stick out of the crowd in the race for global AI supremacy.

A major initiative by the Hungarian government to prepare the economy for digital disruption strives to gather stakeholders from business, research and academia under the AI Coalition. The aim is to draw up a comprehensive strategy for the potential impact and use of artificial intelligence so that local enterprises gain on competitiveness and Hungary could turn into an AI stronghold in Europe.  

As professional leader of the 175 member strong AI Coalition, Gergely Szertics said at an earlier innovation summit that, although it all feels like a race, the idea is to find mutual benefits for players from public and private sectors alike.

“There’s so much happening already, from research to app development to product launches around here, yet these efforts are isolated,” he noted.

However, it is no longer only the research centers of multinational companies like Continental, Bosch or Nokia that promise impact. Targeted research by local players also prepares the ground to go global. Take self-driving car solutions by AImotive, emotions-based ad analysis by RealEyes, or SignAll’s sign language translation toolkit.

Platform technology is being built into products as well. Just think of Indivizo, a disruptive online recruitment service, or iLex, which handles the entire corporate legal process through algorithms. Add early adopter heavyweights such as MOL, Vodafone or Telekom, and you get the idea that AI is, indeed, all around.

“And yet you can’t talk about an ecosystem here, not to mention the lack of visibility or a common effort to build markets based on AI,” Szertics complains. Therefore, a comprehensive country strategy is needed where best practices from abroad can be utilized. Major pillars include education, research, infrastructure building and regulation. Special priority needs to be given to handle the national data asset pool by establishing a National Data Capital Agency.

‘Immense Value’

“We are sitting on vast piles of data that has immense value. It is time to craft structured models for data use and to collect more existing data,” Szertics says. Pushing for more widespread use of AI technology is another objective. A dedicated website, miagyakorlatban.hu (which stands for AI in practice) allows anyone to upload and share their related solutions, thus help the public get acquainted with the latest tech available.

The same purpose is served by an interactive exhibit that was due to take place the day after this issue of the Budapest Business Journal went to print.

Ultinous Zrt.’s AI solution is a great example of the depth of local efforts. The company combines old-fashioned video tech with algorithms to provide the same metrics as user behavior statistics in online stores.  

As commercial director Balázs Horváth said, in offline stores it can precisely track who goes where, and how much time customers spend at a certain spot, so you can easily measure engagement. Other data, such as gender or age, is also available. The system determines age with a four-year degree of accuracy, on average, and re-identification of individuals is also pretty much flawless once they enter the screen. Even mass events pose no bottlenecks.  

“Mass head detection in challenging environments like at festivals is a regular feature, together with blurring faces real time to comply with privacy laws,” Horváth explains.  

One impressive asset of the tech is how it helps cut lines at check-out counters. The AI-powered video analysis predicts whether one will form in the first place. If that is about to happen, cashiers are alerted to open up fresh tills to prevent the lines.  

“In Rossmann, long lines have completely disappeared thanks to our service, whereas short ones were cut by up to 80%. This is what the combination of an existing CCT infrastructure, cloud-based or locally installed servers as well as structured data and real time alerts and insights are capable of,” concludes Horváth.