Editorial Feature

What's Next for a Post-ChatGPT World?

ChatGPT, the generative artificial intelligence (AI) tool provided by OpenAI, has introduced AI to troves of non-specialists around the world since its launch in November 2022. This article examines how generative AIs could change industry, and explores what happens next in a post-ChatGPT world.

Image Credit: CKA/Shutterstock.com

The Viral Rise of ChatGPT

ChatGPT is a generative AI, that is, an AI program that can produce or generate apparently fresh content. Other generative AIs include Dall-E, which produces images from text prompts, and Meta’s Make-A-Video program. ChatGPT is a kind of AI called a large language model (LLM) because it gets information from a large body of written text and uses that to produce new text-based content.

ChatGPT is a free portal for GPT-3, a powerful LLM developed and operated by San Francisco, US-based OpenAI. OpenAI is an AI research company founded in 2015 with backing from major technology investors such as Greg Brockman, Sam Altman, Ilya Sutskever, John Schulman, Wojciech Zaremba, and Elon Musk. Global technology leader Microsoft, which was already invested in the company, recently announced another reported $10 billion investment into OpenAI, motivated in part by the success of ChatGPT.

Shortly after its release in November 2022, OpenAI’s ChatGPT gained one million users. OpenAI is already capitalizing on this popularity with the introduction of a subscription service to a faster, more feature-packed version of ChatGPT for $20 per month.

Part of ChatGPT’s viral success is due to its performance. The program is good at answering questions, generating text content, and classifying textual documents. It does this with a very large capacity, with 175 billion parameters identifying and classifying types of written word.

Other LLMs have been released or are soon to be released. Google is expected to launch its Apprentice Bard LLM imminently, while Meta’s OPT (Open Pretrained Transformer) is already available for industry professionals and researchers to use.

Now, industry experts predict LLMs and other generative AIs will soon be incorporated into general word-processors, data tools, and search engines.

Currently, AI developers are creating more sophisticated LLMs, such as GPT-4, which OpenAI plans to release later this year. There are also dedicated programs in development, such as Med-PaLM7, Google’s clinically-focussed LLM for medical professionals.

How Do Generative AIs Like ChatGPT Work?

Generative AIs are examples of machine learning (ML). These are algorithms that tell a computer to process a lot of information (called training data) in order to gain some “knowledge” or insight from it. In the case of generative AIs, the knowledge gained is how to convincingly mimic the training data with new information or content.

LLMs like ChatGPT are trained on a body of text, which is usually scraped from the internet. Then, the AI algorithm responds to text-based queries with text-based responses that look like the content in its training data.

ChatGPT is a sophisticated AI that performs well due to a few factors: the quantity of training data it has processed, the quality of that data (in terms of it being a broad sweep of content on the internet), and the 175 billion parameters it uses to classify training data, new query inputs, and its own responses.

There are criticisms around the applicability and use of generative AIs, and chief among these is that there is no guarantee that the answers they provide are accurate. They are simply convincing mimicries of the training data.

This unreliability is unavoidable. Generative AI learns and identifies statistical patterns in training data and repeats those patterns to make convincing responses. However, the training data itself is not vetted and may contain inaccuracies, biases, or out-of-date information.

For technical topics where there is less training data available, generative AIs can often return erroneous responses. Sometimes, AIs can even return hateful or offensive responses if their training data contains that kind of content.

OpenAI has attempted to overcome this problem largely by employing human moderators to label text and highlight harmful content. It has been reported that these workers were poorly paid and suffered trauma as a result of their work.

Educators and scientists have also raised concerns about dependence on generative AIs going forward, which could harm students’ ability to develop critical thinking and problem-solving skills. Issues of plagiarism and academic cheating have also been raised.

The comparatively huge computer processing demands of generative AIs also present an environmental problem. Large amounts of energy are required to power servers that can run generative AI tasks, and large quantities of computer hardware (and the rare earth materials used to make it) must be deployed as well.

These downsides may just be growing pains for a relatively immature technology, or they may be unavoidable problems with the technology going forward.

How Will Industry Adopt Generative AI?

Despite the criticisms of generative AI tools and their relative immaturity, industry is already making use of them.

ChatGPT, for example, can be employed to analyze sentiments and opinions expressed in text, such as in product reviews or social media posts. This can give marketing and public relations (PR) experts valuable customer and market insights.

In the digital sector, ChatGPT can be used as the basis of conversational bots and virtual assistants. Bots like this are deployed in education, retail, consulting, and academic research.

Content creation is a key feature of generative AIs, and media organizations already use them to produce content of ranging quality.

Faster, easier text analysis and generation are applicable in nearly every industrial sector. It can help people draft emails, send reports, analyze data, and understand new topics much more quickly than before. This is part of ChatGPT’s runaway success: its embrace across industry by anybody seeking to save time and improve productivity in the office.

CELUS: Using Deep Tech to Accelerate Electronic Engineering

References and Further Reading

Hoyos, A. (2023). Unpacking ChatGPT: The Pros and Cons of AI’s Hottest Language Model. [Online] IE University. Available at: https://www.ie.edu/insights/articles/unpacking-chatgpt-the-pros-and-cons-of-ais-hottest-language-model/. Accessed on 2 March 2023.

Stokel-Walker C., and R. Van Noorden (2023). What ChatGPT and generative AI mean for science. Nature. doi.org/10.1038/d41586-023-00340-6.

Zinkula, J. and A. Mok (2023). 7 ways to use ChatGPT at work to boost your productivity, make your job easier, and save a ton of time. [Online] Business Insider. Available at: https://www.businessinsider.com/how-to-use-chatgpt-at-work-job-save-time-ai-2023-2. Accessed on 2 March 2023.

Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.

Ben Pilkington

Written by

Ben Pilkington

Ben Pilkington is a freelance writer who is interested in society and technology. He enjoys learning how the latest scientific developments can affect us and imagining what will be possible in the future. Since completing graduate studies at Oxford University in 2016, Ben has reported on developments in computer software, the UK technology industry, digital rights and privacy, industrial automation, IoT, AI, additive manufacturing, sustainability, and clean technology.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Pilkington, Ben. (2023, March 15). What's Next for a Post-ChatGPT World?. AZoRobotics. Retrieved on November 24, 2024 from https://www.azorobotics.com/Article.aspx?ArticleID=584.

  • MLA

    Pilkington, Ben. "What's Next for a Post-ChatGPT World?". AZoRobotics. 24 November 2024. <https://www.azorobotics.com/Article.aspx?ArticleID=584>.

  • Chicago

    Pilkington, Ben. "What's Next for a Post-ChatGPT World?". AZoRobotics. https://www.azorobotics.com/Article.aspx?ArticleID=584. (accessed November 24, 2024).

  • Harvard

    Pilkington, Ben. 2023. What's Next for a Post-ChatGPT World?. AZoRobotics, viewed 24 November 2024, https://www.azorobotics.com/Article.aspx?ArticleID=584.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Your comment type
Submit

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.