Uhlík vs Křemík · Czech podcast about artificial intelligence

Machines for us, or instead of us? Where AI helps people and where it replaces them

Episode 2 · 6 February 2026 · 43:04

Listen on: YouTube · Spotify · Apple Podcasts

Artificial intelligence is already changing the job market and the way we learn, and in healthcare it could greatly improve our lives. In the second episode of the Uhlík vs Křemík podcast, host David Salomon asks Adam Hanka whether machines will work for us or instead of us. Hanka, who by his own account has been on the market as an expert for 15 years and sat on the jury of the AI personality of 2025 award, answers that people will win, as long as they use AI well and responsibly.

A more complex world calls for more cooperation

According to Hanka, nobody can keep up with the pace of change. In his view the world is more complex and more technologically connected. One of his favourite managers recently explained to him that the problems we solve today are the ones that have not been solved yet, which makes them harder, so people have to spend more time aligning with each other, building shared understanding and trust. He compares it to medicine: even a simple operation needs several doctors. In the same way, a good AI tool needs a team of people who can communicate and agree on a goal.

How do you stay up to date? Hanka recommends Harari's book Nexus and, above all, trying things yourself. As an example he mentions Google's NotebookLM, where you upload your own sources and have them turned into a podcast or a presentation. A Deep Search type tool once went through 380 sources for him in 18 minutes and 53 seconds. He warns against overload, though: whoever runs eight such searches a day suddenly has 150 pages of text. You need to know what you want, and only then can you ask the AI the right way.

Who holds the power, and why Czechia needs AI in healthcare

The artificial intelligence people know from Google or OpenAI is, according to Hanka, in the hands of American companies. With the Chinese DeepSeek, he cannot say whether it is backed by a private company or by the party and the government in China. He considers European AI regulation to be very much right. In his view America clearly leads the race, Chinese models come second and Europe is third in the AI index he tracks. "So far it looks like the only one who can spoil it for America is America."

He sees a big opportunity in healthcare. According to him, 192 thousand children were born on the territory of the Czech Republic in 1970, while in 1998 the number fell below 100 thousand for the first time. The younger generation will therefore not have enough doctors or carers for their ageing parents. "And we cannot pull this off with human power. We have to pull it with automation, robotics and artificial intelligence."

Fewer juniors and new skills

Replacing people is already happening, Hanka says, especially in large American companies. They hire fewer juniors, because AI can replace their work more easily than the work of an experienced expert. Routine work is at risk too, such as copying data between spreadsheets or answering the same questions in customer support. He mentions predictions that AI could replace 10% of jobs within a few years, and expects people to change jobs more often. His grandmother, by contrast, stayed with the same company for 41 years.

As a defence he recommends communication skills and the basics of mathematics, technology and programming, for example trying to build your own web page in HTML. He considers critical thinking extremely important, and it rests on a reference framework of knowledge, so knowledge is still needed. He explains the fear that AI will make us dumb with the theory of deskilling: people lose skills they no longer need thanks to technology, like monks copying books after the arrival of the printing press. In his view we will use the freed cognitive capacity differently, mainly for filtering information.

The year 2025 and what 2026 will bring

From 2025, Hanka highlights three years since people first saw ChatGPT, and the arrival of the Chinese model DeepSeek with a good ratio of size to performance. With DeepSeek, though, he was not sure whether it arrived at the end of 2024 or at the start of 2025. The overall AI personality of the year was Jan Kavalírek, the commissioner for artificial intelligence at the Ministry of Industry and Trade. Hanka immediately adds that perhaps he no longer is, since the incoming government seems not to want to keep him there, which he considers a great pity. According to Hanka, Kavalírek is also behind Czechia being a hot candidate for building an AI Gigafactory. On the Czech AI scene, he says, people who understand the field in depth work alongside many charlatans who want to cash in on the hype.

In 2026 he expects companies' demand for introducing AI to grow and artificial intelligence to move further into industrial use. The precondition is order in the data. Companies that have cleaned theirs up are, in his words, already making money on their first applications. Asked who will win, carbon or silicon, he answers that artificial intelligence does not compete, so people will always win.

The full conversation is available in episode 2 of the Uhlík vs Křemík podcast.

The episode is in Czech. Czech page with chapters, quotes and the full transcript