Parallel Worlds ChatGPT

rabiakhatun

In March, General Motors announced that it was working on a virtual voice assistant that uses AI models based on ChatGPT. The tool was implemented by Coca-Cola to strengthen its marketing policy. There are dozens of similar business projects.

Let's recall that GPT (Generative Pre-trained Transformer) technologies capable of generating texts have been used at Google since 2017. The ChatGPT model from OpenAI, released in November 2022, created hype due to its user-friendly interface, which led to the mass use of the neural network.

AI systems similar to those developed by content writing service OpenAI and the release of open-source versions can be used to automate many tasks in the economy, making processes more efficient and reducing costs, Olga Zvagolskaya, head of insourcing at ITGLOBAL.COM Group, told RSpectr.

The widespread availability of ChatGPT-style AI systems and the release of open-source versions will be a catalyst for the AI ​​field, Alexey Sergeev, head of machine learning and artificial intelligence practice at Axenix (ex-Accenture), told RSpectr.

Alexander Ushakov, digital product designer at Selecty Group, added in a conversation with RSpectr that




the spread of neural networks should lead to easier public access to AI technologies and the creation of new applications that improve people's lives in many areas - in medicine, logistics, education

Educational institutions are already implementing ChatGPT in the educational process. For example, from March 10 to 12, the IT school GigaSchool held a hackathon, where 30 teams created AI projects.

According to Alexander Ushakov, the entertainment industry will also receive new opportunities with the advent of ChatGPT.

BOTS LIKE MUSHROOMS

ChatGPT gives a weighted average answer based on numerous known data, stated Ivan Lindberg, CEO of OOO Delis. According to him, testing practice shows that the answers are not of high quality.

Ivan Lindberg, Share LLC:

– Such systems should be perceived as demonstration samples, not as professional tools. In terms of information, they do more harm than good.

Any most effective technology is in demand not only by the "right bees". In this regard, undesirable use cases related to fraud, generation of fakes and disinformation are also possible, noted Alexey Sergeev.

Finnish information security company WithSecure published a study in 2023 that claims that

ChatGPT can be successfully used to generate fake news

In December, a group of Western scientists used ChatGPT to write scientific papers. In some reports, the neural network replaced facts about research. In January, ChatGPT misled recruiters from the consulting company Schwa, who mistook the program for a real applicant and wanted to meet with him.

The issue of blurring the boundaries between truth and lies is one of the main challenges of the world of generative networks, Ruslan Akhtyamova, co-founder of Napoleon IT, told RSpectr. According to the expert, a parallel, fully synthesized information world is already being actively created.

People will get used to the fact that when they request information, networks respond correctly and believe all the answers, but many - with the increase in the number of chatbots - will be incorrect. By transmitting information, users will gradually change their beliefs, says Anton Kalabukhov, Marketing Director of Homeapp and Interion. The expert noted the role of bloggers in creating low-quality content.

Anton Kalabukhov, Homeapp and Interion:

– As neural networks grow, they will begin to check facts even less, the entry threshold for new authors online will decrease, and people will write about things they don’t understand at all.