Don't forget risks associated with artificial intelligence

Artificial intelligence is without a doubt phenomenon of today's time. Previous sentence is cliche, but it is true, that almost everyone tried some chatbot powered by artificial intelligence, even just for a curiosity. Today's time is captivating, due to fast creation and implementation of new technologies. In 2015 were AI models just machine learning usable only in limited areas. But, until 2023 were created advanced systems, like generative AI (e.g. ChatGPT), which can create texts, pictures or programs, which change industry sectors in just a few years. Maybe more telling is the speed of adoption of new technologies. While it took television decades to get billions or users, technolgies like smart phones or social sites achieved this milestone just in few years. ChatGPT achieved 100 milion users in just 2 months from the start.

As people, we are captivated by new possibilities and lot of time we forget the risks, and even in a case of artificial intelligence. Some threats like so-called hallucinating of artificial intelligence can seem funny to us. Artificial intelligence can sometimes think of fantastic (but non-existing) recipes. For example, when unnamed model AI got a task to write a recipe for pasta, he answered: "Firstly, cook pasta in milk, then add 500 grams of chocolate, and lastly, sprinkle everything with a pinch of vanilla powder and few pieces of lemon peels." When was different artificial intelligence model asked, to write an essay about evolution of animals, he claimed, that dolphines can play on piano and that bees have their on banks. These hallucinations can occur for example in language models, when they are generating texts based on equations, which they learned, but don't have access to current or verified information. However, this does not prevent people from making decisions based on inaccurate information... If it will be badly written paper by student in school, it won't be that bad, but if the artificial intelligence will help with decision-making of water managers, we can have a big problem.

In 2021 came European commision with a draft of Regulation about artificial intelligence (AIA), which should manage rules for it's development and usage. That's is alright, but it has a catch, it takes effect in 2026! We come back to the speed of technological evolution and the question of how relevant such a regulation will be when at least 5 years have passed since its draft. Will the regulation reflect technological changes and breakthroughs of technology to a practical life? That is a big question to think of.

Make own intern rules for the usage of artificial intelligence and it's connection to intern processes (believe, that even in this moment users use available models of artificial intelligence, to ease their work-load). While non-compliance with basic safety rules can endanger not only private data, but even patents or know-how of organisation. Likewise, uncritical acceptance of artificial intelligence outputs can very easily endanger important decisions.
So if you don't know, how to prepare for aratificial intelligence, how to evaluate it's risks, how to establish intern rules, or just what to do, you can turn to Aricoma. We will help you anytime you need :)