Photo: ChatGPT bot became the most popular application in 2023 (Getty Images) Author: Anna Berezina
A year ago, OpenAI presented ChatGPT. This was a sensation, because ordinary users were able to use artificial intelligence in everyday life. In recent months, the chatbot has become even more advanced, and its popularity is breaking records. At the same time, concerns continue to grow in society due to the rapid development of AI and the difficulty of controlling it.
RBC-Ukraine tells what happened to ChatGPT over the year, where it is used and why there are so many scandals around the bot.
In preparation, the publications of OpenAI, The Telegraph, The New York Times, Forbes, CNBC, Similarweb, the statement of the head of the Ministry of Digital Development Mikhail Fedorov, the publications “Our Kiev”, “Chitomo” were used.
How ChatGPT became a hit in the tech world
ChatGPT debuted on November 30, 2022. The American company OpenAI has opened access to a language model of artificial intelligence in the form of a chatbot for everyone.
In the first 5 days, the number of ChatGPT users reached 1 million, and only two months later – 100 million.
According to Similarweb research, in October 2023, 1.7 billion people used the desktop version of a chatbot or mobile application . The open.ai website has become one of the most visited in the world (21st place in the ranking in November).
Photo: traffic indicators of the most popular sites (similarweb.com)
At the end of 2023, the ChatGPT application on Android was named the best according to users.
The secret to the chatbot's popularity lies in its ability to answer complex questions in simple language . People can ask questions from virtually any area of expertise, and the chatbot responds in a manner that is more reminiscent of a conversation with another person than with digital artificial intelligence technology.
ChatGPT can write poetry and essays, solve mathematical problems, generate program code, translate texts, and give personal advice. In addition, developers are constantly improving its capabilities.
OpenAI's product has provided impetus for other companies to develop their own AI technologies. ChatGPT has competitors (Bard, Claude from Anthropic, Character.AI and CoPilot from Microsoft), but they are inferior to it in popularity.
Bot evolution over the year
As the chatbot was used, its functionality expanded. At first there was only a free version of ChatGPT-3, which provided text responses to queries, and the database was limited to information only until 2021. The bot's responses were often inaccurate, irrelevant, and even fictitious.
However, the latest upgrade (GPT-4 Turbo) allowed access to data from the network for the period until April 2023 . The model is trained on more data.
OpenAI says the updated bot produces fewer incorrect answers and is able to understand “more nuanced instructions” than the previous model. ChatGPT can now “see, hear and speak” . Users can communicate with the bot by voice, exchange images that it processes and responds to.
For example, you can send ChatGPT a photo of a plant and ask it to tell you how to properly care for it.
Another significant innovation is the ability to create your own bots , which does not require programming skills.
“Creating one is as simple as starting a conversation, giving it instructions and additional knowledge, and choosing what it can do, such as searching the Internet, creating images, or analyzing data,” OpenAI explained.
You can share your chatbots online through the “GPT Store.” It is planned to launch in early 2024.
Photo: chatbot is constantly being improved (Getty Images)
It is important to note that for now, GPT-4 Turbo features are only available through a paid subscription (about $20) for the Plus and Enterprise versions. There are also updates for the free version. GPT-3.5 Turbo can process large amounts of information – it can work with a context of 16,000 tokens.
Where is AI technology used?
After the launch of the chatbot, its critics stated that such products with artificial intelligence technologies would soon leave representatives of many professions unemployed , including clerks, marketers, financial advisors, lawyers, accountants, media workers, support specialists, psychologists, and teachers.
Indeed, there are already cases where companies are cutting staff, replacing them with chatbots. Thus, in India, the IT company Dukaan cut 90% of its customer support staff, and the German publication Bild announced plans to introduce AI and fire some workers.
However, at the moment, AI technologies primarily serve as assistants.
Thus, in the UK they are seriously considering the possibility of introducing AI tools into the education system. Teachers are already using artificial intelligence to write letters to parents and check homework. Prime Minister Rishi Sunak believes AI will help low-income citizens gain access to education .
According to him, in the future, each child will have an individual mentor – a digital robot with AI .
As for the ChatGPT bot from OpenAI itself, it was implemented by the following companies:
Expedia , one of the world's most popular travel planning websites, has integrated AI-powered voice assistance into its services.
Microsoft (OpenAI investor) – uses a chatbot in the Bing search engine, allowing users to search and get results through a conversational interface rather than a traditional list of links.
Duolingo , a language learning service, has implemented a ChatGPT model that helps explain the rules to the user, and the user can also practice the language with AI characters.
CocaCola , a non-alcoholic soda manufacturer, uses technology to solve marketing problems (to create personalized advertising texts, images and messages).
The Snapchat app has a built-in bot that can help answer questions or just keep the conversation going.
UK energy supplier Octopus Energy has integrated ChatGPT into its customer service channels and is now responsible for handling 44% of customer queries.
Udacity 's online courses used GPT-4 to create an intelligent virtual instructor that can provide personalized guidance and feedback to students.
The Slack platform has created an application using ChatGPT to manage workflows. Einstein's AI assistant can compose responses, summarize discussions, or conduct external research.
Photo: AI bot is used by successful businesses around the world (Getty Images)
ChatGPT in Ukraine
In April 2023, Minister of Digital Transformation Mikhail Fedorov said that Ukraine was negotiating with OpenAI about introducing the company’s product into government digital services. The Ministry of Digital Development is studying options for integrating ChatGPT, and one of them is a virtual assistant in “Diya” under the working name “Nadiya”.
Oschadbank specialists are considering the introduction of AI, and have already tested the technology this year. ChatGPT managed the bank's social media pages.
In turn, Ukrainian businesses are already actively using the capabilities of ChatGPT. The publication “Our Kiev” wrote about several such companies. Green Forest Academic, a language learning school, uses AI to write textbook materials. The IT company Railsware has implemented ChatGPT to work with code and solve analytical problems; the Youscan company uses a chatbot in its customer support service, for language translation of materials, and working with text.
The first magazine created using AI was published in Ukraine. This is an electronic media called “Colander” about art. The texts for the issue were written by ChatGPT, and the illustrations were created by OpenArt, Midjourney, DALL E.
Scandals and bans: how the world is trying to regulate AI
Skeptics believe that the use of AI can lead to the massive spread of disinformation, manipulation, state propaganda, violations of intellectual property rights, and cyber attacks.
The developers also talk about the risks of artificial intelligence, but believe that we should not abandon further development of the technology, but instead work to normalize its use.
The head of OpenAI Sam Altman himself compared artificial intelligence to nuclear weapons . A few months after launching ChatGPT, he signed a declaration saying that governments should take AI as seriously as the atomic bomb. He has since called for the creation of a global body similar to the International Atomic Energy Agency to address the risks posed by the technology.
At the same time, a scandal erupted around Altman a few weeks ago. The board of directors of OpenAI fired him for allegedly “not always being frank in his communications with the board, which prevented the latter from fulfilling his duties.”
As the NYT wrote, the board's numerous complaints included concerns that Altman was paying too little attention to the dangers posed by AI, and that his resignation was an attempt to seize control of artificial intelligence and slow down the development of technology until there are protective mechanisms in place . However, amid mass protests from OpenAI employees, Altman was reinstated.
Photo: OpenAI head Sam Altman found himself at the center of a scandal (Getty Images)
The problem of AI regulation is very acute at the moment. Legislatures have not kept pace with advances in digital technology , but are trying to take steps to address potential dangers.
US President Joe Biden has issued an executive order on security measures in the field of AI. According to the document, AI companies must report to the government the risks that their systems could help countries or terrorists produce weapons of mass destruction. The order also aims to reduce the threat of “deepfakes” that could influence elections or deceive users.
Japan is developing non-binding guidelines for the technology, while China has imposed restrictions on certain types of AI.
The European Union is promoting a new law on artificial intelligence. It could limit some risky uses of the technology and enforce transparency requirements (all AI-generated content will be labeled).
At the same time, there is already a problem with AI copyright compliance. One of the high-profile cases is a lawsuit against OpenAI by 17 writers , including Jonathan Franzen, John Grisham, George Martin. The lawsuit alleges that OpenAI used the authors' works without their permission or remuneration by downloading them into its artificial intelligence models.
What is the future of AI?
In the future, AI-powered chatbots could evolve into personalized assistants that can cater to a user's specific needs.
Also, more and more so-called copylots will appear – digital assistants built into programs.
According to one of the leaders of OpenAI, Mira Murati, ChatGPT continues to develop and soon a keyboard will not be needed to communicate with the bot.
“You may soon be able to interact with ChatGPT without typing anything,” Murati said in a statement to CNBC.
Also, one of the company’s future goals is to “make bots think for themselves.” ChatGPT should not only provide more accurate information, but also form a certain “chain of thoughts” or reasoning so that the user can understand how the system arrived at exactly such an answer to the request.
According to the OpenAI representative, the development of artificial intelligence technologies brings both opportunities and risks. For example, AI can be an effective teaching tool, and at the same time, personal digital tutors can easily replace teachers.
“No matter what, the revolution is coming. And it will be the public and governments that will determine how and to what extent the AI revolution will affect our lives,” says Murati.
Earlier, OpenAI warned of a development that could destroy humanity. It is an autonomous AI system that can replace a highly qualified specialist with knowledge, skills and experience.
Let us recall that the UK, USA, EU and China signed a joint declaration on the “catastrophic” danger of artificial intelligence for humanity.