Infoshare is one of the most important technological conferences in the country. This year, according to the organizers' data, over 6,000 participants took part in it. The scale, especially looking at the trade fair part and the main stages, is evident. As for the stages, there were 7 of them, with four dedicated to specific industries - marketing and three stages for IT developers. The first observation, looking not only at these three stages but also at the number of exhibitors looking for IT employees, suggests that the information about the IT crisis in our market is exaggerated.
The main theme of the conference was undoubtedly AI, especially generative AI. There were at least a few moments at Infoshare when half of the ongoing lectures were about this topic. It was evident that this year there was one "leading theme" at the conference.
The fact that Artificial Intelligence was the leading theme did not mean that there was a consensus about its rapid development. Despite all the panelists talking about AI emphasizing that the development of generative Artificial Intelligence is the most important trend of the year and, above all, an event that changes the technological scene, not everyone was a technological optimist. Phaedra Boinodiris from IBM stated that the main goal is to ensure that models, whether language or predictive, on which AI is based, are understandable to the user and familiar to them - something that the most popular models do not provide. She also provided three IBM principles regarding AI:
- It should support humans.
- Data and technology utilization outcomes belong to the creator.
- AI must be open and explainable.
These are very important principles in comparison with examples provided by Boinodiris herself - from who is responsible for the inaccurate biography created by Chat GPT, through the model created by the state of Michigan used to estimate fraud when applying for benefits, whose accuracy was less than 10%, to the Spanish algorithm designed to detect the chance of domestic violence recurrence, which was audited after 19 women were killed, whose homes were under investigation. Another panelist, Jowita Michalska, during her panel on technological trends until 2030, stated that due to AI, "we will never think independently again." Therefore, it is crucial to inform about how specific AI models work, what data they are based on, who is responsible for these data, and which social groups are included in them.
Conference on the future - with a few problems
On the other hand, not only thanks to AI, for the first time in several years of attending technological conferences, I felt at Infoshare like at a conference that shows solutions that until recently were the domain of science fiction - from (admittedly somewhat strange) presentations on artificial wombs to a very well-documented lecture on the use and hacking of drones by terrorist groups, to the mentioned panel by Jowita Michalska, which asked questions like "should genetically modified animals be released into the wild." Especially the second day of the conference also included panels that were a significant source of knowledge for start-ups.
However, it was not a conference without problems - many of the panels, especially those on the main stage, were very "marketing-oriented" in the negative sense of the word - promoting very specific company solutions, only lightly touching on the topic - unfortunately, large companies led in this, including Microsoft and Bosch, but they were not the only ones. This is probably the first time that panels with such little value for the listener appeared at this event.
However, in my opinion, Infoshare, thanks to networking, the number of available companies, a wide range of topics, or even accompanying conference events, certainly remains one of the most important events on the start-up scene. The level is also maintained by the startup competition, which truly selects the most interesting solutions at an early stage of development, not only on the Polish scene.