It is now possible, after almost two years since the beginning of the artificial intelligence (AI) boom, to assess its impact more realistically on the economy, including the labour market as well as its future. This future, as well as challenges and new opportunities, was discussed in Karpacz by those participating in the debate ‘Artificial intelligence – Quo vadis?’.
As Katarzyna Baliga-Nicholson, manager at Sano – Centre for Personalised Computing Medicine, pointed out, nowadays pragmatism supersedes thinking about the threats posed by AI. We already see that not so many stations but specific, often cumbersome tasks performed by humans are being automated with the help of AI. “The promise of technology entering areas unwanted (by humans - ed.) is being fulfilled,” the Sano expert pointed out.
According to Marcin Chrząszcz of Akamai Technologies Poland, although many senior managers see AI as the solution to all problems, it still takes a lot of data and a lot of time for artificial intelligence to become a real competition for humans. For example, it takes more time for humans today to catch and fix bugs (errors) made by ChatGPT when generating code than to write it themselves. This is because data quality is crucial in AI development, and for now it is a major challenge.
This challenge was also noted by Professor Ryszard Tadeusiewicz of the AGH University of Science and Technology in Kraków, who reminded us that the machine learning (ML) algorithms used in AI yield better results if they have a lot of data. The quality and reliability of this data is a big problem, as it may not be correct. The Internet resources used by AI include both reliable and questionable information, as well as deliberate lies.
“So far we have not been able to solve this problem. We do not know how to check the quality of the data that we so merrily put into ML algorithms. There is a risk that AI may threaten us because people lead it astray,” Professor Tadeusiewicz emphasised.
