As enthusiastic that Sam Altman or Elon Musk are when talking about the benefits of AI, not everyone agrees. Rare is the day in which new studies do not criticize this technology. From those who point to students who use it to those who directly say that people return to people. Now experts point to university.
Or, specifically, to university research. Apparently, The arrival of artificial intelligence to academic environments is a serious danger: that they lose prestige. So much that there is no lack of those who argue that it is necessary to take measures, and take them soon. Otherwise, it will be increasingly complicated to discern what is true of what not.
AI, a «poison» for university research
Since always, university studies, research carried out in these educational centers have had considerable interest. When a job carried the firm of Oxford, Harvard or the one that was, deserved the credibility of the public. But irruption everywhere in artificial intelligence and tools such as chatgpt can change the thing enough.
That, at least, is what defends a new article published in The Australian. According to the teqsa (something like the Australian regulatory agency), Artificial intelligence could be «poisoning» many university research. Mainly, because they question the authenticity of many of the data that are now used to be able to carry them out.
It is what they call «Data Poisoning.» That is to say, When models like chatgpt give for good information that is not really verifiedwhich is not what is said to be precise or, simply, the AI has taken it out of the sleeve. Everyone already knows their inaccuracies, or even their hallucinations. Problems that, experts believe, would already be affecting many conclusions.
If the trend is consolidated and prosperous, The result would be none other than largely reducing the credibility of scientific findings. And as a possible solution, they only see one exit: that students are honest and recognize when they are using AI. In this way, his works will be more caught with tweezers, so to speak. Of course, that is to say, and another very different.
The risks of artificial intelligence, put on the table
The investigation carried out by the Australian Teqsa does not end there. In the same way, he has also denounced a problem that has been on the table for a long time. Their experts consider that AI and Chatgpt are a risk to privacy and intellectual property rights. Very important aspects that do not always have as much as it should.
If this work is right, everything indicates that the quality of university findings could be compromised. Therefore, they also insist that it will be increasingly necessary to monitor exhaustive of everything that is done. Otherwise, AI can complicate a lot to know when something is true and when not. And if science is not accurate, then it is difficult to know what it would.
Know How we work in NoticiasVE.
Tags: studies, artificial intelligence