The deployment of Artificial Intelligence (AI) and Machine Learning (ML) in recent years have greatly impacted our social and economic lives in unimaginable ways. Individuals, businesses and industries are relying on the predictive tools of AI and ML (data integration and analysis) to enhance their lives and operations. A survey conducted by McKinsey and Company in 2022 showed that AI adoption has doubled over the past five years, and investment in AI is fast increasing.
Particularly, generative AI is using technology to change the narrative on what used to be the preserve of the human mind – ingenuity – through the creation of content from generated data. A proof of the usefulness and uniqueness of generative AI is the infamous ‘ChatGPT’ – which has stirred the world with its engineered language feedback responses to user prompts, outperforming the functions of a traditional search engine.
Following its launch, the technology has witnessed massive global adoption – over 1 million users in less than 5 days – surpassing any technology adoption milestone. The growing craze for ‘ChatGPT’ comes with its ability to provide personalised, contextual and specific answers in a manner equal to answers provided by humans. This adoption-rate, according to experts, has sparked global discussions around the technology’s potential disruption of Google’s US$200billion search business.
What is ChatGPT?
‘ChatGPT’ is a chatbot built on OpenAI’s Generative Pretrained Transformer 3 (GPT-3) model, which functions as a Large Language Model (LLM). GPT-3, a third version of the advanced language model developed by OpenAI, is an ultra-modern language-processing AI model with a wide range of applications – including language translation and language modelling – which can generate answers to questions posed by its users in a manner similar to answers provided by humans.
In its relatively short deployment period, GPT-3 already uses over 175 billion parameters to recognise and form sentences and word groups.
Large Language Models (LLMs) are AI tools that can read, translate texts, summarise and predict words like a human would – just like the autocomplete function in most text or word-based interfaces. However, ChatGPT is a buildup on regular LLMs as it incorporates the Reinforcement Learning with Human Learning techniques by using human modalities to train machines to provide specific answers to users’ prompts. The integration with other learning techniques is the result of LLMs’ inability to fully comprehend users’ needs to provide useful predictions.
Unlike a simple chatbot, ChatGPT is specifically trained to understand the human intent in a question to provide helpful answers. Thus, it can explain, programme and argue with human intelligence and efficiency. As part of its functionality, ChatGPT is a calibration of the GPT-3.5 model – which was trained on huge amounts of data about code and information from the Internet, including sources like Reddit discussions, to help ChatGPT learn dialogue and attain a human style of responding.
Use cases and impacts of ChatGPT
Within the short period of its launch, ChatGPT has brought about significant changes to our digital interactions and our work performance. It has maximised the use and analysis of data by individuals, businesses and industries for optimal outputs. Below are some areas which have been immensely impacted by the ChatGPT technology:
This chat-based language model has stirred up almost every aspect of the teaching and learning cycle, and it has been met largely with both amazement and fear by educators since discovering its potential to disrupt the learning experiences, especially, of students. Admittedly, students are always looking for easy routes to learning; but the efficiency of this language model is quite worrying – especially for the fear that students may never learn to write. For a model that can deliver apt and well-written essays and solve complex assignments, we can bet on the over-reliance of students on its capabilities.
Early in the year, the New York City Department of Education reportedly blocked access to ChatGPT on all department devices and networks – with its justification being the negative impact it was going to have on students’ learning. However, despite the general concerns which have been raised, the ease this technology brings to students’ work cannot be over-emphasised.
While some educators are opting for incorporating the language model into the educational system to maximise its usefulness in the teaching and learning process, others have criticised its use and adoption – as, according to them, it is ‘killing’ creativity among students. Currently, there are ongoing engagements by stakeholders to ascertain the adoption and use of this technology by teachers and students.
In Ghana, stakeholders in the educational sector must begin the engagement, discussion and design of reforms which incorporate the usefulness of this technology into our education system and foster its positive adaptation.
- Search engines
With the chatbot’s ability to promptly generate natural language text accurately, ChatGPT has facilitated the generation of answers to searches made by its users. Due to its functionality, experts’ reviews have suggested that the technology could pose a threat to Google – our long-cherished search engine. However, there is a school of thought which holds the view that the ChatGPT technology is only a tool and not a service or a product; hence, it can be built into search engines like Google to enhance their data output.
The opportunity to incorporate this technology into existing search engines may be the reason Microsoft has recently agreed to invest over US$10billion in the bid to improve its search engine ‘Bing’ with a scheduled roll-out in March this year. This integration can potentially threaten Google’s dominance as the most patronised search engine.
However, one limitation that the technology faces is the range of information it can provide to its users upon request. Currently, its data output only extends to the year 2021, and it is unable to provide data beyond this date. An example is the chatbot’s inability to name the country that won the recently-ended World Cup.
Although this is a concern, some have expressed the view that this can be resolved in subsequent iterations. Among the many issues identified with its usage is the concern of copyright acknowledgments. This is so because the chatbot synthesises information from different sources, hence its inability to provide absolute references for the data it provides. In this light, emerging conversations are highlighting how the chatbot’s usage may infringe copyrights. Hopefully, OpenAI will proffer a solution in an updated version of this revolutionary technology.
- Marketing and customer service
In our world where businesses are looking to provide cutting-edge market solutions in a very competitive market, this technology has the potential to generate personalised content for a marketing campaign – tailored emails and social media posts for different types of customers. As one of its key features, ChatGPT has the ability to provide quick and detailed responses in conversations – automated responses to customer inquiries – as well as to increase customer experiences by addressing the main pain-points for customers.
It is anticipated that ChatGPT will become an indispensable aspect of digital marketing and customer service through the provision of more personalised and helpful responses to customers.
It is interesting how the use of ChatGPT is also impacting our social and entertainment lives. The use of the language model knows no bounds – ranging from writing threads for social media pages, to writing novels to designing games. The chatbot has been trained on texts in books, articles and websites that have been structured in the supervised learning process so as to write books, make-up songs, write codes, etc.
This is of little wonder as OpenAI’s DALL-E, an art generation-based model, has been used to generate art and designs in all manner of shapes and colours based on specifications communicated to it by users. A recent improvement in its capabilities is its ability to provide a summary of a movie. As much as this may sound unbelievable, it is true.
One of the major ways ChatGPT has had an impact on finance is through its ability to handle complex financial queries and provide accurate, reliable responses. This has made it an invaluable resource for financial advisors, who can use it to assist clients speedily and efficiently.
Also, the technology has been used as a model for new-generation chatbots which assist customers with tasks such as tracking investments, checking account balances etc., all through simple, natural conversations. Its usefulness in the finance industry transcends customer service. It has been employed to analyse financial data and help prospective investors make prudent investment decisions.
Institutions and firms are leveraging the technology’s machine learning capabilities to identify and predict trends, so as to streamline their solutions to suit the diversified needs of their customers.
Indeed, technology has affected the way we work; and following the onboarding of this new technology, our work-life is bound for major transformations. An obvious way in which ChatGPT is impacting the job market is through automation – the advancement of technology has led to the deployment of technological solutions to ‘compete’ with humans on the performance of manual tasks – mostly data integration and analysis services.
There is a harrowing view that the adoption and use of this chatbot is the beginning of the end for some professionals. Thus, its operation can possibly lead to the displacement of workers whose roles can be efficiently and easily performed by this chatbot – including data analysts and customer service representatives.
Nonetheless, the point has to be made that these language-generation models are not completely capable of replacing all manual tasks; and they are most effective when used to complement human capabilities rather than replace them. Many tasks still require human creativity, judgment and critical-thinking skills. The development and subsequent maintenance of these generative models can rather create new job opportunities in fields such as software engineering and data science.
Individuals and businesses will have to view work differently in the nearest future, as ChatGPT has come to stay.
The advancement of AI and related technology is improving the quality of our everyday lives: from effectively performing routine and complicated tasks to making life simpler, safer and more efficient. The real and potential benefits of the revolutionary AI-powered chatbot ‘ChatGPT’ cannot be discounted, and should be leveraged by individuals and businesses to boost effectiveness and productivity.
>>>the writer is a Trainee Associate at Sustineri Attorneys PRUC (www.sustineriattorneys.com). Cecilia has an interest in Financial Technology and Innovations, Startups and SMEs, Commercial Transactions, Company Law and Contracts, as well as ADR. She welcomes views on this article and is reachable at [email protected]