Dr. Pradeep Mahapatra
After one year completion of the public launch of Generative Artificial Intelligence (GenAI), the earlier days of 2024 witnessed enough speculation of possible impact of the new technology on everyday life of the people worldwide. Following footsteps of ChatGPT, Large Language Models (LLM) like Google Bard, BingChat, ClaudeLLM are available free of charge, while Jasper solicits subscription for use.
In addition, applications like Adobe Firefly, Bedrock, BloombergGPT, Dall-E2, Eighthfold.ai, Grammarly Go, HR Signal, Microsoft-365-Copilot, Midjourney, Retrain.ai, Salesforce-Einstein-GPT, Soundbite.ai, SlackGPT, Tome etc. are available for professional use. Experts in the field claim that dominance of GenAI in various professions will result in fundamental changes in their operations. Journalism is one of the forerunners in the transformation scenario.
LLMs such as ChatGPT can assist journalists in tasks of content research for writing news stories, generating ideas, drafting and summarizing. Further, it can help for interview preparation by providing background information on various topics. Generally journalists spend more time to attend to these routine tasks which can be completed in a moment by utilising the technology. It can enhance productivity streamlining the writing process.
Since GenAI is undergoing an ‘evolution process’, it is difficult to predict the possible total transformational picture. A section of experts confessed during 2023 that 85 percent of the jobs that will exist in 2030 have not been invented yet. Progress of in the field of the new technology doubles in six months which makes long term estimation not feasible.
In the post-pandemic ecosystem journalism faced failures in holding the ground with business, trust and supremacy in the society. New developments made life of professional journalists miserable. In such an atmosphere, adoption of a new technology poses as yet another problem. Journalism has confronted several technological innovations during past seven decades after the Second World War. Journalists are aware that those who can not adopt to the changes will turn mis-match to the evolving system. However, many are not aware how they can prepare themselves for using LLMs.
ChatGPT, Bard, Bing and Claude applications are available free of cost to everybody for downloads in Android computing instruments such as smartphones and laptops. For the first time in the technological history people worldwide could access to an emerging skill in the real time. It created a democratic digital order. During earlier days for adoption of older technologies, the user had to undergo extensive training. But LLMs do not need any such formal training. LLMs are simple chatbots and whenever somebody asks a question or directs a task, the bot attends to it with immediate result. Thus, the users have to learn how to ask questions or give directions ‘Prompt’ to the bot with personal endeavours.
The three features of LLMs distinguish themselves to be different from the learning process of earlier technologies. To learn steps to be taken by professional journalists, a question was asked to the ChatGPT. As per the answer presented by the bot, “Journalists can learn to leverage ChatGPT effectively by familiarizing themselves with its capabilities through hands-on practice. Exploring the model’s responses, understanding how it interprets queries, and experimenting with different prompts can enhance their understanding. Additionally, staying updated on new features or improvements in language models and considering ethical considerations in AI use are essential aspects of learning to use ChatGPT in journalistic context.”
Many practioners post their experiences on use of LLMs in different platforms. A few specialised websites like JournalismAI publish case studies which helps to understand the practical aspect of the technology. Resources on the subject are also available in open.ai blog and random posts in medium.com etc.
In ChatGPTs understanding, “The effectiveness of using ChatGPT often relies on the prompts given to the model”. In a 10 point scale, importance of prompting demands 9 points. “A well-crafted and contextually relevant prompt significantly influences the quality and relevance of the model’s responses. Experimenting with different prompts and understanding how to guide the model’s output through effective inputs is crucial for maximizing the utility of ChatGPT in various tasks.”
“The answering mechanism of ChatGPT involves a combination of pre-training and fine-tuning. Initially, the model is pre-trained on a large dataset containing parts of the internet. It learns grammar, facts, reasoning abilities, and context from diverse sources. This phase helps the model understand language patterns and relationship. After pre-training the model is fine-tuned on a narrower dataset with the help of human reviewers. When a user inputs a prompt or question, the model generates a response based on its learned knowledge and patterns. The generated response is influenced by the input prompt and the context it provides.”
On behalf of ‘World Association of Newspapers and News Publishers’ (WAN-INFRA), a report on status of the contemporary news industry entitled “World Press Trends Report 2024” was published during second part of January 2024. The report contained inputs received from 175 senior newsroom executives spread over 60 nations. It claimed that 87 percent of news platforms are prepared to utilize AI in their organisation. But 72 percent had not ready for implementation till the end of 2023. Thus, though AI may not be used in news rooms by 2024, it may spread during the following years. It is being pointed out that lack of capacity to use, ignorance about the technology and speeder technological change pose hindrance for early use of AI in journalism. In an environment clad with problems, capacity building on AI on individual level seems to be good option for professional journalists.
(English translation of the original Odia newsletter by the author circulated on February 9, 2024. https://pmjournalism.substack.com/p/dc3. It is an open-access content, free for translation and reproduction)
Dr. Pradeep Mahapatra is a retired faculty of Journalism, Berhampur University, Odisha.https://about.me/pradeepmahapatra
References:
Dollins, Mark & Elizabeth Ballow. Artificial Intelligence & the communication. www.northstar.comms.com 2023
World Press Trends Report 2024 : Twipe’s top 5 insights
