According to an analysis by Bloomberg, the Internet and social networking sites have started to see a number of sites that publish news in the form of breaking news, dealing with several political, social and economic issues. number of followers, who have interacted with this news, but they don’t know that all of this The stories these sites publish are all generated from popular AI tools such as OpenAI’s ChatGPT and the new Alphabet, Bard.
The company specializing in the evaluation of the news “NewsGuard” has tracked these sites, revealing the existence of 49 different sites, all of which publish information written entirely thanks to artificial intelligence, some of which is based on real information. published in newspapers and magazines, and others that were made from misleading anonymous YouTube content or videos.
By closely following these sites, NewsGuard found information on one of them reporting the death of US President Joe Biden, and that his deputy Camilla Harris is in charge of the country, and followed another site reporting the death of thousands of soldiers. during the war in Ukraine, citing an unconfirmed video clip, posted on YouTube.
Research has shown that most of these sites are “content farms”, which are low-quality websites run by anonymous sources that post advertisements, spread around the world and published in multiple languages, including English. English, Portuguese, Tagalog and Thai.
By following these sites, some of them were able to generate income through advertisements by posting content from another reality, but others tried to build an audience on social networks, such as a publication site of biographies of celebrities whose Facebook page has 124,000 followers, and thanks to the increase in the number of followers, they can generate income through advertisements.
The real problem here lies in the large number of such news that it is difficult for ordinary people to distinguish whether it is true or false, and given that we are only at the beginning of the era of intelligence services which provide written text services, the problem is likely to get much worse if these services are not controlled.
NewsGuard CEO says report shows importance for companies launching AI services to train their bots not to help write fake news and use them for such negative things .
actions to contain the problem
The list of risks that have arisen since the emergence of artificial intelligence robots is growing rapidly day by day, and it indicates only one thing, which is the need for government and official regulations governing the work. in this sector, which has experienced rapid growth. and a huge development in a short time, which has countless people worried. Many experts, led by Elon Musk, met with US Senate Democratic Majority Leader Chuck Schumer and other lawmakers last week to discuss measures to regulate artificial intelligence. .
Musk has been among the harshest critics of the current rapid pace of AI development and signed an open letter last March with the Future of Life Institute calling on AI labs to stop for a while. 6 months on the training of more powerful artificial intelligence systems than the “GPT-4” program. ‘, referring to the latest version of ‘ChatGPT’.
Reasons and dimensions
Future fears in the AI ​​industry prompted us on business website Sky News Arabia to discuss the issue with an expert in AI and robotic programming, Islam Shatnawi, who said that AI is an algorithm or numerical model that can learn from experience. , and those experiences come from all forms of data, be it text or images. Or visual, like video, and so he can create models from which he can learn. If taught experiences in illegal or erroneous areas, unfortunately, there will be a way to learn and generate knowledge about those illegal or erroneous matters. Certainly the subject of fabricated news and directing public opinion through it is considered one of the negative aspects of artificial intelligence, in the same way that a person uses his expertise to do illegal or bad things. things, making it a concern for many, especially with the spread of things like so-called “deep forgery”, and the emergence of fake accounts on communication sites behind which thousands of bots are hiding . that they are natural people to direct certain content or information and publish it to a large number of people. All of these examples show how creative artificial intelligence can be in doing these bad things, but we also have to talk about the so-called counter-artificial intelligence, which detects if this fake news, traces it and draws a tree sources to expose deceptive artificial intelligence and significantly reduce its dangers
responsibility and solutions
Asked about the responsibility for these dangers and problems, is it the responsibility of the companies that launched these services? Or the people who use it? Or governments that have not yet regulated the sector? Shatnawi responds by saying that the responsibility is shared by all those mentioned. Countries must establish laws to control ethical standards and set levels of transparency and impartiality for the algorithms that run the artificial intelligence services that operate there, and conduct a periodic audit of companies that provide artificial intelligence services and set standards. Through it, you assess how well these companies’ algorithms meet the ethical standards you have set.
Companies should also ensure data protection and privacy and monitor any adverse aspects that may result from the use of their AI services, and periodically review the transparency and independence of the algorithms on which their services run. Artificial intelligence are based by applying standards like the IEEE initiative to ensure security standards and level of ethics in algorithms.
Shatnawi adds that the user also has a responsibility: if he discovers an erroneous aspect of the artificial intelligence systems, he must point it out and reveal it, and that companies encourage individuals to accept this disclosure through prizes, For example.
Islam Shatnawi focused on a very important issue, namely that the problems that appear in artificial intelligence systems are based on the learning sources of the algorithms of these systems, and unfortunately these sources do not always fit into the framework from ethical or correct sources, and to sum up the subject, what is the algorithm on which artificial intelligence systems work? It is only a mirror of our data and not an independent output from these systems.
Will the laws solve the crisis?
And on whether enforcing the laws governing the AI ​​industry will solve the aforementioned problems, Islam Shatnaoui told Sky News Arabia that there is no complete and integrated system, and that there is there will always be errors, and therefore lessons should be learned, as the purpose of these systems is to reduce risk and error. As much as possible through review, continuous updating and quality checking of resources learning algorithms and their level of compliance with government standards, these procedures do not guarantee that errors and problems will disappear 100%, but they will ensure access to the optimal version and the least problems in the country who will apply it.
Read the Latest World News Today on The Eastern Herald.
