Counter gas for AI

Now that the initial amazement about what AI (artificial intelligence) can do and the benefits that this entails, has subsided, it now seems that there is more attention for the disadvantages.
It started with a letter from a number of scientists and tech figures such as Elon Musk and Apple co-founder Steve Wozniak, calling for a temporary pause in the development of systems more powerful than ChatGPT-4.
Subsequently, Italy temporarily banned ChatGPT due to privacy concerns about processing personal information of Italian users. And the United States now also wants stricter testing before AI is released, although it is not clear what role politics should play in this. Over the next two months, the possibility of AI audits and other measures that could address concerns about these new systems will be explored.

AI is increasingly used

Meanwhile, in recent years, the chatbot has already become a standard part of websites, such as Amazon, to help people with returns and refunds. Last year, OpenAI released an advanced version, ChatGPT. ChatGPT uses deep learning techniques to produce text based on input. But instead of programmers having to enter data, large amounts of data are entered into the system. The program then uses that information to train itself to "understand" information.
However, there is a caveat: the program does not 'know' what it produces. In fact, this "understanding" is nothing more than the logical sequence of words that often occur in combination with each other. Based on the large amounts of text it has received, the program sees that the words 'the' and 'table' are almost always used as a combination and then puts them one after the other.
But this is just the beginning. The authors of the aforementioned open letter point out that current AI systems have an intelligence equivalent to that of humans. They wonder whether we should allow ourselves to be inundated with propaganda and untruth, whether we should automate all jobs, and whether we should develop machines that can eventually outperform or replace us?
The letter writers argue that such decisions should not be left to “unelected technology leaders.” They say companies should develop safety protocols for advanced AI design and development, which can be overseen by independent third-party experts.
In short, the discussion on this topic will continue for a while.