Six surprising facts about ChatGPT

Nowadays, people are talking about ChatGPT, an artificial intelligence chatbot, and now it’s time to really understand what’s going on.

ChatGPT is an artificial intelligence chat tool developed by OpenAI. It can write poetry and prose like humans, and even write software code.

Conversations center around topics of surprise, fear and warning – surprise at the power of artificial intelligence, fear at deception and possible replacement by “robots”, and fear at outsourcing creativity A warning about the dangers posed by machines to humans.

In fact, ChatGP is not as surprising as most media reports, less scary, and requires less warning. ChatGPT is an artificial intelligence language model developed by San Francisco-based company OpenAI that can converse with people through text and generate a wide range of content on request.

Generative artificial intelligence is by far the most dominant technology trend of this decade. So, here’s how to better understand it by knowing 6 facts about ChatGPT that people may not have heard of.

1. ChatGPT is not the only artificial intelligence chatbot with this function

The special thing about ChatGPT is that it is an artificial intelligence chatbot for public use. There are other AI projects out there that are just as powerful or even more powerful, but many people have never used them.

Major global technology companies (especially Google, Meta and others), as well as some startups and universities, have developed generative artificial intelligence tools that are at least as powerful as ChatGPT. In some cases, OpenAI's competitors have released limited versions of artificial intelligence, and some countries have not made similar technology available for public use.

Meta company released a chatbot called Blenderbot before ChatGPT. But the bot is so discreet in conducting reviews that users find it boring to use.

The caution with which these AI systems are audited stems from unfortunate encounters in the real world. Microsoft released a chatbot called "Tay" a few years ago that was trained on a large number of social media posts and sourced content from extremists and venters online, so it posted racist and vitriolic messages. The remarks embarrassed Microsoft.

Google fired an engineer last year after he publicly claimed that Google's LaMDA (Language Model for Dialogue Applications) was aware.

ChatGPT has been a hot topic in the tech world for the past two months, thanks to OpenAI’s bold release of its tool to the public, allowing the chatbot to address some (but not all) controversial topics. While other tech companies are more cautious, they are rushing to launch related products now that ChatGPT is getting so much attention.

According to reports, Google has released internal "code red" to accelerate the public use of its artificial intelligence products. The company plans to launch more than 20 new artificial intelligence tools this year, and Baidu will release a chatbot similar to ChatGPT in March this year.

Meanwhile, a website called Futurepedia lists hundreds of artificial intelligence tools people can try, many of which are serious competitors to ChatGPT.

While ChatGPT is not the only powerful artificial intelligence system, its boldness and usability are unique.

2. Microsoft is the largest investor in OpenAI

The for-profit OpenAI LP company that developed ChatGPT is a subsidiary of the non-profit company OpenAI Incorporated, but OpenAI Incorporated is not really " Owns "OpenAI LP" company.

Microsoft is the largest investor in OpenAI LP. Microsoft plans to invest another US$10 billion in the company, giving it a 49% stake. The company's parent company, OpenAI Incorporated, owns only 2% of the shares, and other investors will share the remaining 49%.

Microsoft is also actively working to build ChatGPT capabilities into Bing, Azure, PowerPoint, Outlook and other products.

Ultimately, Microsoft controls OpenAI.

3. Most criticisms of ChatGPT are not related to its technology

The three biggest concerns about the widespread use of ChatGPT are: (1) students will use it to cheat; (2) Will be used for unethical purposes such as plagiarism and social engineering attacks; (3) False information will sometimes be provided.

One can ignore the first two concerns because generative AI systems that work like ChatGPT will have tools that can detect their content. AI chatbots like ChatGPT are inevitably subject to detection and will only get better. People want to avoid their unethical use.

In any case, human moral shortcomings are not the fault of AI tools.

The third worry is wrong. When ChatGPT or other chatbots offer false claims, it’s only because the training data set contains false information that hasn’t been caught by moderation tools, including the current public “beta version.” Because bashing ChatGPT for bad data is like bashing a movie for watching a bad movie because it misses the point.

In the future, businesses will use generative AI on their own data or on data that has been thoroughly vetted. In other words, technology is one thing, data is another.

Imagine an agency like the CIA feeding all of its intelligence reports, political analysis, and records of millions of phone calls into an AI system like ChatGPT. The potential for this technology to gain insights quickly is unimaginable and staggering.

4. ChatGPT was developed by humans

It is easy to think of the use of text-based artificial intelligence as interacting with computers rather than humans. In fact, projects like ChatGPT source human-produced content through programs written by developers, which are then reviewed and prioritized by humans. Therefore, it contains human errors, biases and conclusions.

5. ChatGPT requires user skills to realize its full potential

In fact, all text input generation artificial intelligence tools require skilled input to achieve maximum effect. Prompt Engineering is the art and science of knowing how to talk to each AI tool. This new skill will be similar to SEO in developing and using knowledge about which words produce which results. .

"Prompt Engineering" is an emerging profession, and this is also the field that most people will get involved in.

6. Some industries have begun to rely on ChatGPT

artificial intelligence systems. Some types of professional writing are "robotic", automatic, and formulaic. Some studies have found that artificial intelligence can do this tedious work faster and cheaper.

The industry that may adopt ChatGPT is the real estate industry. When listing a property, you can use parameters to describe a property for sale and list some data (square feet, total area, number of bedrooms and bathrooms, price, etc.) and descriptive and factual terms, such as "Renovated", "Open Concept", "Restaurant" etc.

Real estate agents now simply feed all their data into ChatGPT, and the artificial intelligence system writes an overview of their properties.

Real estate is not the only industry adopting ChatGPT, the industry is just ahead of other industries, and many industries will rely entirely on artificial intelligence tools in the future.

The launch of ChatGPT has attracted everyone’s attention, and in fact, it is now in the early stages of generative artificial intelligence. Understanding the reality of what's happening with this technology is what everyone needs to care about.

The above is the detailed content of Six surprising facts about ChatGPT. For more information, please follow other related articles on the PHP Chinese website!