Chat-GPT was launched by OpenAI in November 2022 but rapidly grew to over 100 million users by January 2023. But tougher rules for chatbots like chat-GPT are now awaited in Europe since the EU Parliament in early June 2023 voted on new rules for artificial intelligence. But what exactly is chatGPT?
Chat-GPT is an AI chatbot. You can ask it anything and get an answer. But unlike a simpler voice assistant such as Siri, chatGPT is built on a so-called LLM (Large Language Model). These are trained on huge amounts of internet information, from which they generate completely new answers rather than just picking up ones already stored. They are not built for a specific service like a regular chatbot but are much smarter.
EASY TO USE
Using chatGPT is easy, as all you have to do is type in your question. Not satisfied with the answer? Give the bot additional instructions. Since it keeps track of previous questions, you can refine from there instead of starting over again. Chats are then saved in conversations in the sidebar until you delete them. At the same time, the answers cannot be completely trusted. If chatGPT lacks the knowledge, it can fill in the gaps with incorrect information – it ”hallucinates”. According to OpenAI chatGPT also has ”limited knowledge of world events after 2021.”
MANY ARE CONCERNED
In addition to making mistakes, many are concerned about what this human-like AI could mean for the future of the internet – so much so that Twitter and Tesla CEO Elon Musk, along with other tech leaders in March 2023, called for an immediate pause in AI development.
Italy became the first country in Europe to ban the use of chatGTP. And many companies – such as Apple and Samsung – have prohibited their employees from using chatGTP. In Sweden, the Swedish National Agency for Education (Skolverket) recommended to avoid home exams entirely to protect against plagiarism and abuse. Namely, essay writing for students is one of the most obvious examples of where chatGPT can become a problem.
WHO OWNS THE COPYRIGHT?
Already there are hundreds of books published on Amazon where the bot is listed as the author – and certainly a whole lot where the correct author is not specified. But who actually owns the copyright to content created by chatGPT is a question open to debate. Some argue that generative AI ”steals” the content it’s been trained on – something that has become even more controversial in the world of AI art. Regardless, the topic of copyright law regarding generative AI needs to be adressed, especially since copyright law, as it currently stands, only protects content created by humans.
”If you are presenting it as your own work and you have done it with the help of an AI, then it is up to you to make sure that what the AI is doing is legal. If we claim that a person is the author, then that person will also be responsible. It’s about rights and obligations – they go hand in hand,” says Eleonora Rosati, professor of intellectual property law to DI Digital.