“Groundbreaking”, “mind-blowing”, and “magical” are just a few of the numberless adjectives techies around the world have used to describe ChatGPT, a chatbot created by OpenAI. Apart from bringing human-like interactions, the bot has also opened doors for more innovation in conversational AI ( artificial intelligence), say industry leaders.
ChatGPT is the latest update of OpenAI’s language generation model Generative Pretrained Transformer 3 (GPT-3). The free-to-use prototype has been trained on a massive amount of text data and uses deep learning algorithms to generate natural language text similar to human writing. Using a dialogue format, it can answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.
When asked what is exceptional about it, ChatGPT writes, “I am one of the largest and most powerful language generation models currently available, with 175 billion parameters, which allows me to generate highly detailed and nuanced responses.”
However, its functionality is not limited to writing essays and poems. The generative model can even build HTML code to create a full website within 30 seconds. Experts on conversational AI said this makes it a significant development.
“The launch of ChatGPT opens up opportunities for us in two major ways. First, it has the potential to create mass awareness about Conversational AI and automated virtual assistants which can lead to increased adoption expanding the market for us. Second, since it is open source, we can always integrate it into our product and make it more powerful,” said Rashi Gupta, Chief Data Scientist and Co-founder of REZO.AI, an automation company based out of Noida.
Rezo helps the contact centers of companies like Maruti Suzuki, Tata AIG, and ACT Fibernet, resolve customer queries more quickly.
“The new bot can handle broadly a lot of B2C use cases like weather forecast, recommendations for travel location, what to pack, DIY searches, etc. It's a great development and will add users to get more precise consolidated responses without spending hours on searches,” Gupta added.
For the business process management industry this could be good news, of course, this is still the early days of ChatGPT. By 2026, conversational AI deployments at contact centres will reduce agent labour costs by $80 billion, says a Gartner report. It projects that one in 10 agent interactions will be automated by 2026, compared with 1.6 per cent now.
Worldwide end-user spending on conversational AI solutions at contact centres is projected to reach $1.99 billion this year.
ChatGPT’s user count crossed the one million mark within five days of its launch on November 30. On the other hand, inFeedo – the developer of employee engagement platforms has been using the bot, right from its previous version GPT-3, to improve its solutions.
“Our SaaS platform Amber uses the GPT3 language model to summarize, acknowledge and take various other quick actions by understanding and categorizing what employees are saying, at scale. We also plan to use this when talking to employees and collecting feedback by having human-like conversations,” said Tanmaya Jain, Founder & Chief Executive Officer (CEO) of inFeedo.
Jain said that generative AI can have multiple use cases in HR from recruiting, and employee onboarding to employee experience. “It opens new doors for industries heavily dependent on customer service, content curation, and more. I am especially excited to see the possibilities that ChatGPT will open up in the employee experience and engagement industry, as it will enable us to drill down into tons of data and understand what is useful and actionable in seconds,” he said.
Open AI offers open-source application programming interfaces (APIs) with GPT-3, which performs a variety of natural language tasks, Codex, which translates natural language to code, and DALL·E, which creates and edits original images.
However, industry players caution that it is still early days. Beerud Sheth, Co-founder, and CEO, Gupshup said the ChatGPT will become “very impactful” for end users though it might currently be a playing tool.
“Even with its exceptional understanding, ChatGPT has more generalized use cases as it lacks specific understanding about a business, industry, or company. Companies working in the conversational AI space will have to adapt it based on the verticals they want to use for retail, BFSI, automotive, or anything else,” Sheth said.
The prototype has several limitations too. As per OpenAI’s website, it may occasionally generate incorrect information, could sometimes produce harmful instructions or biased content, and has limited knowledge about events that happened after 2021.
“ChatGPT is incredibly limited but good enough at some things to create a misleading impression of greatness,” the CEO of OpenAI Sam Altman said in a Tweet. He said the tool is not good enough to be relied upon for factual queries.
The innovator adds, “It's a mistake to be relying on it for anything important right now. It’s a preview of progress; we have lots of work to do on robustness and truthfulness… We will work hard to improve!”