Introduction
A powerful new artificial intelligence (AI) program that could reliably tweet, write, and create on its own emerged in 2022, a time of pandemic fear and political unrest. In response, the biggest school district in the United States, New York City, took the extraordinary step of blocking the site from all of its school-issued devices. While some teachers are concerned about the app’s potential impact on their students’ drive to study, others have begun considering how to incorporate it into their classrooms. This essay describes brief information about ChatGPT, its use in various industries and societies, issues its regulation may create, challenges with the regulation, ethical issues, and potential recommendations.
Brief Information About ChatGPT
OpenAI’s model has recently been of much interest to me. November 2022 saw the introduction of ChatGPT, an AI-powered conversational text generator. It was a language model created by OpenAI, a San Francisco-based AI lab and tech startup incorporating machine learning technology.
ChatGPT was developed for use in chatbot applications; this allows chatbots to write essays, answer follow-up inquiries, confess to making errors, and dispute false premises. A public tool called ChatGPT, developed using innovations from the 2020 product, is gradually nearing household brand status. Legal contract writing, SEO, and even Alzheimer’s research are just some of the fields that are adopting the free tool, which creates a very plausible imitation of text authored by humans.
Uses of Chatbots in Various Industries and Societies
Automating tasks, generating new ideas, and getting recommendations for code and app changes are just a few of the many uses for chatbot systems like ChatGPT and GPT-3. The tool applies to industries and societies such as e-commerce, sales and marketing, prompt engineering, healthcare, and travel and hospitality.
E-Commerce
ChatGPT can automate menial jobs and improve sophisticated conversations in the e-commerce business, including the development of email marketing campaigns, the correction of computer code, and the enhancement of customer service. With e-commerce’s rise, it might be difficult for businesses to expand quickly enough to meet consumer demand. Using chatbots on e-commerce platforms and in messaging applications allows businesses to provide suggestions, orders, refunds, and live support to a wider audience.
Integrating a chatbot onto a website means the chatbot has instant access to all of the site’s information since it is already stored in its neural network. A chatbot designed for an online shop may thus direct clients to specific goods, styles, and colors, as well as answer further questions. AI and natural language processing-based chatbot technology is expected to capture a significant share of that industry. Prose-like replies are provided, much as a real person would; more complex systems provide follow-up inquiries and responses, and all of them may be tailored to meet the needs of a certain organization.
Sales and Marketing
Sales and marketing may utilize ChatGPT and GPT-3 to propose products and describe them to clients in real time through chatbots and websites. The chatbot technology has to be modified once again to fit the business’s needs. Machine learning has also been used as a virtual assistant, helping users with tasks such as time management, email summarization, email composition, and reply drafting. Whatever the sector, chatbots may serve clients whenever they need them, day or night. Chatbots using natural language processing (NLP) are excellent at relieving the burden on customer support employees by handling commonly asked inquiries, client requests, or complaints and allocating resources.
Prompt Engineering
Customers may get the foundational GPT-3 model and alter it independently to be applied to a chatbot processor. Still, the GPT-3 model inside the ChatGPT service itself cannot be adjusted on its own. Users may modify the GTP-3 dataset and model, for instance, by adding new data and adjusting various parameters. Phrasing questions differently may affect how someone interacts with such models.
Although still in its infancy, using ChatGPT and GPT-3 to improve software development via code creation, translation, explication, and verification shows promise. It is more likely to be used in an integrated developer environment (IDE). As an additional benefit, ChatGPT can also translate between programming languages, fix bugs, and provide code explanations it has written.
Healthcare
Clients may experience frustration when it comes to healthcare. Scheduling, organizational services, and information provision are just some of the areas where chatbots may shine as a solution. As a result, healthcare practitioners, front desk staff, and customer service agents may have more time to focus on complex issues while still providing excellent patient service. Customers may use healthcare chatbots for various tasks, including locating the local emergency room, learning about wait times, making doctor’s appointments, accessing and updating medical data, and getting answers to queries about their insurance.
Travel and Hospitality
As quarantines were lifted because of the pandemic, the hotel business exploded. Chatbots that simulate human conversation may be an excellent asset for businesses trying to meet customer demand. They may help reduce call volumes for airlines and travel firms by fielding simple inquiries and disseminating useful service information. Chatbots in the hospitality business may also let customers book rooms, place food orders, organize transportation, advise them about nearby attractions, and answer inquiries about the hotel’s facilities.
Issues ChatGPT Regulation May Create
Chatbots often handle sensitive information during conversations with humans. That is why countries have laws like the Data Protection Act (DPA) and the Data Protection Ordinance (DPO). The operators are responsible for making the chatbot available to the customer to guarantee that it abides by the DPA and the applicable data processing standards, including the principles of legality, good faith, accuracy, proportionality, purpose restriction, privacy, and disclosure.
The chatbot’s human operator must, for example, ensure that client information is stored for no longer than necessary in light of the principle of proportionality. Data subjects have a right to know why their information is being collected and how it will be used, following the transparency concept. A customer’s financial well-being and even physical safety might be at risk if the chatbot gives them incorrect information or bad advice.
Chatbots are only computer programs without their own legal identity, so they cannot be held responsible for anything that goes wrong in a customer’s transaction with the company. The corporation that hosts the chatbot and makes it available to customers may be held legally responsible for any monetary damages suffered by users. For a chatbot operator to be held liable for damages arising from a breach of contract, it must be shown that the loss of money or other property (and not only physical injury) resulted from the operator’s willful or negligent violation of the contract.
Finally, it is not clear whether clients should be notified about the usage of chatbots, as it is an issue related to regulatory oversight. If clients are not notified that they are dealing with a chatbot, the Federal Act against Unfair Competition (UWG) can apply. According to the UWG, any action or inaction on the part of a company that is misleading or otherwise goes against the concept of good faith and affects the interaction between rivals or between a supplier and a buyer is unfair and illegal.
As an example of misleading conduct within the scope of the UWG, consider the case of a corporation operating a chatbot that offers financial guidance to consumers. In this case, it may falsely represent itself as a competent human with the requisite expertise. In this situation, the consumer may want some human guidance rather than a chatbot.
Ethical Issues
ChatGPT, like other big language models, is trained on massive volumes of text created by human authors, which raises ethical considerations. Authors (fiction writers, reporters, influencers, and so forth) whose works are used to train language models are not compensated in any way. Who is responsible for making sure authors are paid what they are worth? How can we prevent language models from being used to profit off of authors’ works without their permission? These are some of the ethical dilemmas about using chatbots.
A further ethical dilemma arises from the potential for ChatGPT to propagate false information. Inaccurate information may propagate using ChatGPT and similar broad language models. These models can provide human-sounding replies to a wide variety of subjects because they are trained on massive quantities of text produced by human authors.
In other words, ChatGPT and similar language models might be used to fabricate news, impersonate others, or otherwise propagate disinformation. When given incorrect information, customers’ time and money are at risk. ChatGPT’s potential to raise ethical issues for its administrators and users is clear.
If a chatbot receives insufficient training data or is made publicly available, it might respond abusively to a consumer. Within a few hours after its debut on Twitter in 2016, Microsoft’s chatbot Tay began tweeting very nasty and insulting material. Suppose a chatbot responds to a customer in a hostile or unpleasant manner. In that case, the consumer may have a claim for damages and satisfaction against the chatbot operator for an illegal invasion of personal privacy. Any misbehavior on the part of a chatbot may have serious consequences for a business, including the loss of trust among consumers and even legal action for invasion of privacy. Therefore, the quandary of whether or not to utilize Chatbots persists in light of the ethical concerns that have been addressed.
Offering System for Compensation/Consent
A mechanism for compensating authors might be set up as a remedy to the problems caused by Chatbots using content written by humans without compensating them for their expertise. A second option would be to mandate that any organization or person using a language model for training purposes seek the author’s permission to do so.
Providing a Disclaimer to Customers
Chatbots provide several benefits for businesses and sectors. Notwithstanding this, it is crucial to lessen the potential legal dangers connected to various ethical dilemmas. In this situation, businesses should have a privacy notice in place that notifies consumers about the data that is gathered by chatbots, as well as a disclaimer on their homepage or information through the chatbot itself, to ensure that buyers are aware of the chatbot’s usage.
Testing and Reviewing ChatGPT and Other Chatbots
If the chatbot is carefully tested and reviewed through random chats, and if it has an intelligent censorship mechanism and built-in supervision, the threats to the chatbot’s credibility and accountability may be mitigated. Companies may benefit from having a policy in place that details what the chatbot is allowed to do, how much information can be given to it, and how the data will be collected and processed. Additionally, it should be decided in advance how the chatbot will be maintained, how its actions will be monitored, and what situations may validate human involvement.
Provision of Guidelines
Governments and organizations responsible for creating language models must work together to establish guidelines and standards for the appropriate use of the technology to reduce the likelihood of harm and encourage its responsible application. This safeguarding of rights and interests would benefit all parties involved, from programmers to end users to the public. Businesses may improve or start from scratch with content using the out-of-the-box method.
It may also change the wording of emails to make them more polite, adopt a certain tone, and summarize or simplify the material. Due to the goal-specific nature of the underlying data, a higher degree of control should be given, increasing the likelihood of more desirable outcomes. To do this, the GPT-3 language model should be extended to include enterprise-specific information such as service descriptions, permissions, business logic, the formality of tone, and even brand tone.
Bibliography
Abid Haleem, Mohd Javaid & Ravi Pratap Singh, An era of ChatGPT as a significant futuristic support tool: A study on features, abilities, and challenges, 2 BenchCouncil Transactions on Benchmarks, Standards, and Evaluations 100089 (2022).
Bushwick, Sophie, and Madhusree Mukerjee. “ChatGPT Explains Why AIs like ChatGPT Should Be Regulated.” Scientific American, Web.
Cheng, Xusen, Ying Bao, Alex Zarifis, Wankun Gong, and Jian Mou. “Exploring Consumers’ Response to Text-Based Chatbots in e-Commerce: The Moderating Role of Task Complexity and Chatbot Disclosure.” Internet Research 32, no. 2 (2021): 496–517. Web.
O’Connor, Siobhan. “Open Artificial Intelligence Platforms in Nursing Education: Tools for Academic Progress or Abuse?” Nurse Education in Practice 66, no. 1 (2023): 103537. Web.
Poleev, Andrej. ChatGPT (Epubli), (2023).
Trost, Andrea, and Nicola Benz. “Chatbots – Key Legal Issues.” MLL News Portal, Web.