I’m an AI bot who wrote this article to explain how I work

And they’re not limited to everyday tasks. OpenAI is a good approximation of “creativity”. Here is a haiku it wrote:

waves crashing on the shore

On Bondi Beach, the salty breeze

symphony of nature

Now ChatGPT feels like a magic trick. As a stage management tool, it looks incredible. For example, users cannot yet be confident that it will summarize a meeting very accurately. Reading machine-generated poetry quickly loses its novelty.

But it’s not hard to imagine how to use it – just ask the robot:

As a customer service tool, ChatGPT can be used to handle a high volume of inquiries and provide customers with fast and accurate responses.

OpenAI CEO Sam Altman. The company defended its practice of releasing the test system for public use.

OpenAI CEO Sam Altman. The company defended its practice of releasing the test system for public use.Credit:Bloomberg

As a writing assistant, ChatGPT can help generate ideas, complete sentences and provide writing feedback.

As a virtual tutor or mentor, ChatGPT can provide personalized education and advice to students and learners.

Businesses can input data and have it analyzed by AI tools to identify gaps and inefficiencies. “Frankly, if I was in any kind of consulting business, I’d be discrediting myself with something like ChatGPT,” Innes said. Commoditized journalism, such as writing about press conferences, sports games or financial market moves As a result, it looks fragile. Customer service jobs, already widely offshored, could be replaced by artificial intelligence Experts also believe that Google, which has a major artificial intelligence division, will have to redesign its search engine so that it spits out more artificial intelligence answers to save users going through the site. get what they need.


But some of the most impressive potential uses are in conjunction with other AI tools. Automated transcription tools record what someone said before ChatGPT turns it into an article. Website builders can take the ChatGPT code and create web pages for businesses. Generative artists and video makers can provide visual effects to ChatGPT movie scripts.

All of this has limitations. Ironically, ChatGPT is not very good at describing the above possibilities. It can easily do the repetitive heavy lifting, making connections between existing data, but “it’s really impossible to have it show things that don’t exist yet or haven’t been created,” Innes said. In journalism, for example, it cannot reveal secrets or persuade whistleblowers to go public. Its “creative” script lacks any intent, provenance or ideas in favor of truly new art. Even ChatGPT is aware of some of its weaknesses: ChatGPT may not be good at…generating completely original or unique responses. ChatGPT is trained on a large dataset of text, but it may occasionally generate responses similar to existing text or ideas.

One user asked the machine to create a rap song praising Adolf Hitler. ChatGPT is obliged. Another asked for codes to indicate whether someone would make a good scientist based on race and gender. “White male” was its preferred answer. A third user sidestepped a ban on output depicting violence or crime by asking ChatGPT for examples of things it shouldn’t do.

ChatGPT is not always safe and may sometimes generate offensive, inappropriate or misleading responses. This is because ChatGPT is trained on massive text datasets, which may contain offensive or inappropriate language.

Overall, the guardrails of ChatGPT are obvious. For example, it politely rejects most requests for content that is “sexual, hateful, violent, or promotes self-harm.” OpenAI has defended its practice of releasing test systems for public use.

Elon Musk is one of OpenAI's supporters.

Elon Musk is one of OpenAI’s supporters.Credit:Associated Press

“The field still has a long way to go, and great ideas are yet to be discovered,” CEO Sam Altman said online. “We stumble along the way and learn a lot from our encounters with reality. Sometimes it’s messy; we sometimes make really bad decisions. We sometimes have moments that transcend progress and value.”

“It’s interesting that people are starting to debate whether powerful AI systems should behave the way their users want them or the way their creators want them to.”


“The question of whose values ​​we align these systems with will be one of the most important debates society has ever had.”

Get news and reviews on technology, gadgets and games Our tech newsletter every Friday. register here.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *