AI Prompting: A Solvable Issue
[Article written on June 2, 2023, by Jeremy Lamri with the support of the Open AI GPT-4 algorithm, contributing around 10%. Images created with DALL-E 2 Experimental, all rights reserved, 2023].
In the coming months, generative Artificial Intelligence (AI) will play an increasing role in communication and critical thinking support. However, this reality comes with some challenges, including inequalities in users’ ability to prompt these AI systems effectively. Several factors contribute to these disparities, such as access to tools, linguistic, cultural, and technological competencies. At a time when technology should be the driving force behind democratizing knowledge and skills, understanding and overcoming these disparities become crucial.
The rapid advancements in AI and Natural Language Processing (NLP) technology provide new perspectives and opportunities to address this issue. By enhancing human-machine communication systems, these innovations could not only make generative AI more user-friendly but also mitigate inequalities among different user groups.
What is prompting?
Before delving into the topic’s core, it is essential to define “prompting.” In the context of artificial intelligence and communication with machines, prompting refers to initiating, guiding, or activating a response from a generative AI system.
A generative AI is characterized by its ability to comprehend and produce textual, audio, or visual content, often based on past data examples. Users interact with these systems by providing instructions, questions, or indications — called “prompts” — that help the AI generate relevant and coherent responses or creations. A typical example is the use of ChatGPT, where text queries are the prompts.
In essence, prompting is the process by which a user communicates with a generative AI system by sending messages, requesting information or actions, to receive the desired response or assistance. The way prompts are given directly influences the quality of the response. For example, an AI, like ChatGPT, does not really answer your questions; it predicts the most probable words following your query, and then the following ones, until it forms complete sentences — similarly to Microsoft Word or some smartphones, which suggest the continuation of the sentence you are typing.
Hence, the more accurate and clear in context the prompt is, the easier it will be for the AI to predict suitable words and phrases as a response, based on its vast knowledge database. Beyond the algorithm’s quality and its training data, the generated response quality relies directly on the user’s ability to formulate clear and effective prompts and navigate interactions with generative AI systems. Prompting becomes a major inequality factor as not everyone has the same capacity to pose a query.
Current Inequalities in prompting generative AI
Inequalities in prompting generative AI are a significant challenge that not only hinders equitable access to these innovations but also curbs democratization and adoption. Causes and consequences of these disparities are manifold.
Today, access to technology is not yet universal, mainly due to the digital divide, which translates into a gap between populations with easy access to information technology and the internet and those who are deprived of it. Consequently, some individuals or social groups lack financial resources, infrastructure, or knowledge required to fully leverage generative AI.
Generative AI is predominantly developed by companies located in industrialized countries, consequently emphasizing their languages and cultures. Users unfamiliar with these languages or not sharing the same cultural references may encounter difficulties in prompting AI systems, leading to barriers and misunderstandings.
Individuals have varying communication skills, often due to their education, personality, or professional experience. These variations can create inequalities in prompting generative AI abilities, widening differences between users.
The gap between those who can effectively prompt generative AI and those who cannot runs the risk of deepening social divisions. Individuals with these skills will have access to benefits like higher-paying jobs, learning opportunities, and improved personal growth, while others might be left behind in this technological revolution.
Inequalities in prompting generative AI lead to the exclusion of certain social, cultural, and linguistical groups. Such exclusions can reinforce stereotypes and discrimination while marginalizing these groups in an increasingly connected and technology-dependent world.
Promising technological advancements for better prompt understanding
Innovations in Artificial Intelligence (AI) and Natural Language Processing (NLP) offer promising solutions to improve interactions between users and generative AI systems. Digital assistants like Siri, Google Assistant, and Alexa are perfect examples. These technologies allow users to ask questions and provide instructions using their own words, facilitating more accessible and fluid communication for all.
Moreover, automatic speech recognition (ASR) has improved significantly in recent years. ASR systems can now accurately understand and transcribe a wide range of accents and dialects. This advancement helps eliminate comprehension and consideration issues regarding different communication styles in generative AI systems.
To address inequalities linked to prompting generative AI effectively, adopting interdisciplinary approaches is critical. This means combining sociology, philosophy, and technical sciences’ perspectives and skills to better understand users’ specific needs and expectations.
For instance, sociologists can help analyze cultural and social trends to design more inclusive systems. Philosophers can contribute with their ethics and logic expertise to ensure fair and equitable interactions. Lastly, engineers and IT experts work on developing AI technologies by considering sociologists’ and philosophers’ perspectives to create more suitable and efficient systems.
For technological advancements in prompting generative AI to effectively reduce inequalities, it is imperative to make these technologies accessible to as many people as possible. Several initiatives are already underway to ensure more democratic access to technological tools:
- Lowering costs: Technological progress and increasing demand have led to a decrease in smart devices and AI-related services’ prices, enabling more people to use and master these technologies.
- Reducing entry barriers: Efforts are made to simplify and create user-friendly interfaces, making tools accessible even for those less technologically adept.
- Online educational resources availability: Various free or low-cost resources are available online to help users familiarize themselves with new technologies and improve their prompting skills.
The positive impact on reducing inequalities related to generative AI prompting
AI and NLP innovations rely on increasingly sophisticated approaches to adapt to users’ specific needs. These tools can now better analyze and comprehend prompts, regardless of their quality, accounting for cultural, linguistic, and contextual variations.
For example, deep learning algorithms allow these systems to analyze and interpret nuances in expressions, jargon, and regional dialects, facilitating a more accurate understanding of users’ requests and intentions.
Establishing inclusive and accessible learning environments to allow users to master generative AI prompting is essential. Training and education programs must be tailored to cater to diverse audiences, taking cultural, linguistic, and socioeconomic differences into account.
For instance, free or low-cost online training could be offered, covering topics such as optimal prompting system use and effective communication with generative AI systems. Furthermore, developing easily understandable tutorials and guides contributes to making learning accessible to a broader audience.
Investing in advanced and inclusive AI and NLP technology development can contribute to fostering a more equitable social and economic dynamic. Technology-related sectors offer numerous employment and growth opportunities for marginalized groups, particularly those excluded because of their limited communication skills or access to cutting-edge technologies.
Moreover, by promoting inclusion and enabling more users to prompt generative AI effectively, the exchange of ideas, innovation, and collaboration between various populations and cultural communities is facilitated.
Conclusion
To better understand the issues surrounding prompting generative AI, it is important to consider the advancements in artificial intelligence and natural language processing. Thanks to these developments, we have seen a proliferation of tools like digital assistants and automatic speech recognition, which greatly facilitate communication with machines. These technologies aim not only to improve work processes but also to create closer connections between users and generative AI, catering to individual needs and offering tailored solutions to each user.
By building on these advancements, we can envision a future where prompting-related inequalities among generative AI users are significantly reduced or even eliminated. Interdisciplinary approaches blending sociology, philosophy, and technical sciences allow us to better identify users’ precise needs and develop tools that meet their expectations, regardless of their technology expertise level. Moreover, the democratization of access to technological tools encourages more people to participate in this digital revolution.
Education is a critical lever to enable large-scale, efficient, conscious, and responsible adoption and facilitate understanding of emerging challenges. Learning must be inclusive and accessible to everyone, regardless of their social, cultural, or linguistic backgrounds.
A more equitable future in prompting generative AI is within reach if we continue to think about tomorrow with tomorrow’s codes, rather than our current frameworks. Technology will rapidly evolve to interpret even the poorest prompts more effectively. New technology should be expected to bridge society’s gaps instead of widening them. However, on the other hand, technological advances alone cannot overcome inequalities: organizations play a major role in promoting equal opportunities and access for everyone.
[Article written on June 2, 2023, by Jeremy Lamri with the support of the Open AI GPT-4 algorithm, contributing around 10%. Images created with DALL-E 2 Experimental, all rights reserved, 2023].
— —
If you are interested in the combination of web 3 and HR, I invite you to subscribe to the dedicated newsletter that I keep writing on the subject, and to read the articles that I have written on the topic:
