VOA科学技术2023--ChatGPT Expands with Voice and Image Abilities

ChatGPT Expands with Voice and Image Abilities

The artificial intelligence (AI) tool ChatGPT has added new abilities that include voice and image recognition.

The changes permit some users to directly ask ChatGPT questions and receive voice answers. In addition, the tool can recognize images and provide information about what is in them.

ChatGPT is a chatbot, a computer-powered tool designed to interact smoothly¹ with humans and perform high-level writing. The technology is also known as "generative AI."

The creator of ChatGPT, OpenAI, announced the tool's latest additions, or upgrades, this week. Currently, the new voice and image upgrades are only available to users of ChatGPT's Plus and Enterprise services.

ChatGPT's main service, called GPT-3.5, is free for all users. ChatGPT Plus costs $20 per month. The Enterprise service is designed for individual companies, with costs tied to the services used by the business.

OpenAI explained that ChatGPT Plus and Enterprise users would be able to use the voice and image additions over the next two weeks. The upgraded tools would be made available to other groups of users, including developers, "soon after." The voice and image upgrades will also be added to devices using the iOS and Android systems in the near future.

The company said ChatGPT's new voice control is designed to provide a natural way for users to communicate with the AI tool in a way that is similar to speaking with a human. But it noted² the chatbot can do more than answer questions. It can also tell a story to children or provide detailed³ instructions for making or building something.

Users can choose different voices they want the chatbot to use. The company said it worked closely with professional voice actors to make the interactions more realistic and personal.

The voice interaction ability of the ChatGPT upgrade already exists in many voice assistance systems. These include Amazon's Alexa, Alphabet's Google Assistant, Apple's Siri, and others. American software maker⁴ Microsoft added voice controls to its new ChatGPT-powered Bing search engine earlier this year.

Another notable change to the ChatGPT tool is image recognition. This permits users to upload a photo to the system and then get information about what is contained in the picture.

For example, the company says a user could take a picture of what is currently available in their refrigerator. After entering the photo into ChatGPT, the tool could suggest dinner possibilities based on what the person has. The system could also provide step-by-step instructions for preparing the meal.

Another example given is a parent who might take a picture of a child's math problem and then seek advice on how to explain to the child how to solve it. There is even a way for users to mark areas of the image – for example with a circle – to get more specific information or help with that element.

Along with its announcement, OpenAI issued another warning about how its ChatGPT tool can easily get things wrong. It noted that because the system is trained using massive amounts of publicly available information, it can return results that are false, outdated⁵ or discriminatory.

The company urged all its users to watch out for misinformation and to attempt to verify the information provided by chatbots.

OpenAI announced its AI technology is also being used by the digital music service Spotify. ChatGPT is being used to power a system designed to permit Spotify podcasters to translate their shows into different languages. The translations are completed in the podcasters' own voice in an effort to make them sound more natural, OpenAI said.

Spotify says the first languages to be added in the coming weeks will be Spanish, French and German.

Words in This Story

instructions –n. (pl.) directions for carrying out a task or on how to do something

refrigerator – n. a device used to keep things (such as food and drinks) cold

verify – v. to confirm or prove that something is true

translate – v. to change from one language to another

1 smoothly   

adv.平滑地，顺利地，流利地，流畅地

参考例句：

• The workmen are very cooperative,so the work goes on smoothly.工人们十分合作，所以工作进展顺利。
• Just change one or two words and the sentence will read smoothly.这句话只要动一两个字就顺了。

2 noted   

adj.著名的，知名的

参考例句：

• The local hotel is noted for its good table.当地的那家酒店以餐食精美而著称。
• Jim is noted for arriving late for work.吉姆上班迟到出了名。

3 detailed   

adj.详细的，详尽的，极注意细节的，完全的

参考例句：

• He had made a detailed study of the terrain.他对地形作了缜密的研究。
• A detailed list of our publications is available on request.我们的出版物有一份详细的目录备索。

4 maker   

n.制造者，制造商

参考例句：

• He is a trouble maker，You must be distant with him.他是个捣蛋鬼，你不要跟他在一起。
• A cabinet maker must be a master craftsman.家具木工必须是技艺高超的手艺人。

5 outdated   

adj.旧式的，落伍的，过时的；v.使过时

参考例句：

• That list of addresses is outdated,many have changed.那个通讯录已经没用了，许多地址已经改了。
• Many of us conform to the outdated customs laid down by our forebears.我们许多人都遵循祖先立下的过时习俗。