ChatGPT, OpenAI’s latest creation has had more than a million users since its release last month, with requests to craft code, write a bedtime story and draw up a plan for a new app among the early use cases. Some experts suggest it has an important role to play in enterprise, so Tech Monitor went to the source and asked ChatGPT “what should a CIO know about you?”
ChatGPT is the latest model to be spun out of OpenAI’s Large Language Model GPT-3 and allows users to converse with it using natural language dialogue. It is trained on a vast amount of data that gives it the ability to respond to a variety of requests.
It isn’t always accurate. Coding question website Stack Overflow has banned AI-generated answers due to a high volume of incorrect but plausible responses. In a statement, Stack Overflow said: “Because the average rate of getting correct answers from ChatGPT is too low, the posting of answers created by ChatGPT is substantially harmful to the site and users.”
OpenAI says the tool is still “in development” and they are making changes including training the model based on user feedback, so incorrect code and responses help improve the system if users explain where it went wrong. It is only available for beta testing and evaluation now, but it is expected OpenAI will open API access early next year. This will allow companies to develop products based on the software, which could include coding, optimisation, and call centre tools.
What can ChatGPT do for CIOs?
To find out what ChatGPT thinks its potential is within the enterprise sector Tech Monitor asked it a number of questions, including a request to generate a mock report on managing legacy data, which roles will be taken by AI and what risks should be considered when implementing a large language model like GPT-3.
“As a language model, I can assist with natural language processing tasks and can be used to generate human-like text, but I am not capable of making decisions or performing any actions,” ChatGPT declared when asked what a CIO should know about the tool. “My purpose is to assist users in generating text and to help with natural language processing tasks.”
It outlined a number of areas where it could directly benefit enterprise tech leaders including generating reports, summaries and written comment, helping with data analysis, generating responses to frequently asked questions, providing natural language processing for applications or systems and helping with automation from a simple human-language query.
Content from our partners
“It’s important to note that, as a large language model, I am not capable of making decisions or performing any actions,” ChatGPT warned at the end of a list of potential use cases.
“I can only generate responses based on the information that I have been trained on, and I do not have access to the internet or any external information. I am designed to assist users in generating human-like text and to help with natural language processing tasks.”
To get more specific responses you have to be very specific in your request. For example, Tech Monitor asked ChatGPT to generate an example report that could be used to outline how to make the best of legacy data within an organisation.
It suggested “implementing a strategy for managing and leveraging this data. This could involve migrating legacy data to a modern, scalable platform that can support current and future data management needs. It may also involve implementing data cleaning and enrichment processes to improve the quality and value of the legacy data”.
The top five benefits, according to ChatGPT, of deploying a tool like this within the enterprise environment are: Improved efficiency, automation of tasks, better support for customers, improved data analysis and enhanced natural language processing capabilities.
Will ChatGPT take human jobs?
With those capabilities in mind, Tech Monitor also asked ChatGPT about the possibility of it taking on the job of CIO in the future, leaving humans out of the equation – but it doesn’t think that is a likely scenario anytime soon because “AI and automation technologies are currently not capable of performing” the various tasks delegated to a CIO.
It listed these as including “managing an organisation’s information technology and systems” as well as “planning and implementing technology strategies, managing budgets and resources, and ensuring the security and reliability of the organization’s systems and data”.
The main reason it doesn’t think AI will take those jobs in the near future is that AI “are not able to make strategic decisions or manage complex systems and processes” but warned, “it is important for CIOs to stay informed about the latest developments in AI and automation technologies, and to consider how these technologies can be used to improve the efficiency and effectiveness of their organisation’s technology systems.”
Even after being quizzed further on how it views the role of the CIO, including breaking it down into what the role does and doesn’t involve, ChatGPT was confident humans would keep that job. What a relief.
But that doesn’t mean all jobs within the enterprise sector are safe from automation. When pressed on jobs that could be lost to artificial intelligence in the future ChatGPT gave a list. This included data entry, data processing, customer service, data analysis, financial analysis, quality control, inventory management, supply chain management, fraud detection and transcription.
It explained that “these jobs often involve performing the same tasks over and over, and are well-suited to automation by AI” with jobs involving creativity, judgement or expertise the least likely to be replaced by AI – including R&D, strategy and management.
“It’s important to note that AI is not currently capable of replacing all of the tasks and responsibilities associated with these jobs,” the chatbot cautioned. “Instead, AI and automation technologies can assist with certain tasks and processes, and can help to improve efficiency and productivity.”
ChatGPT is still a “research project” with no indication of when or to what extent it will be made publicly available or accessible to enterprise clients. There is no information on future costs or use cases but it is just one of a number of large language models in development and so this type of natural language solution is likely to play a large part in business in the not-too-distant future.
ChatGPT for enterprise – the analyst viewpoint
Shelly Kramer, principal analyst at Futurum Research described the possibilities presented for enterprise use of large language models as “exciting and daunting”, particularly during this nascent stage of development.
“As it relates to the ability to use a single model for multiple tasks and also used for situations where a model can be trained with minimal data and learn over time, even potentially recognising things it hasn’t specifically been trained on, I see much opportunity,” Kramer said.
“That said, today, this is an incredibly expensive undertaking and one that only the big, well-funded players in the space can afford to do. They are, however, doing the early stage heavy lifting here, using LLM and exploring/proving out test cases by building out proofs of concepts and prototyping.”
While it might be expensive today, Kramer said it is likely they will get bigger, more powerful and exponentially less expensive to train in a fairly short period of time.
“The ability to use AI-generated text for negative, or even evil, purposes is as possible as it is to use it for good,” she adds. “These models can and do learn biases, and context is often challenging. It’s challenging for humans, why wouldn’t it be challenging for LLM? We are in a stage that I usually describe as ‘we don’t know what we don’t know, until we know’. I’m cautiously optimistic, but also pragmatically concerned.”
