Generative Artificial Intelligence and LLMs (Large Language Models) have emerged abruptly into society’s consciousness coinciding with the public availability of tools such as ChatGPT. The amazement that these capabilities have produced and the easy access to experimentation have triggered debate about the possibilities and impact on our work, society, and lives.
The debate is highly relevant, as we notice that the speed at which Generative AI is progressing and being applied is extremely high, where product and algorithm updates are constantly evolving and exceeding the performance of previous versions by several orders of magnitude.
This is only the beginning
Generative AI is a subfield of artificial intelligence that focuses on the creation of algorithms and systems capable of generating new and creative content.LLMs are systems that use deep neural networks to process and generate human language. Examining the short history of Generative Artificial Intelligence/LLMs developments (to which we will dedicate a post soon) we realize the exponential nature of their progression:
- The first productive Generative AI models and LLMs available on large digital platforms emerged less than 10 years ago. This period also saw the creation of the main startups specializing in building and training algorithms.
- Twitter took almost two years (730 days) to reach one million users, Facebook did it within 300 days; ChatGPT did the same within five days of its launch; within three months it had surpassed 100 million users.
This unprecedented speed shows the importance of early awareness of new developments in Generative AI before they become mainstream in order to adapt successfully to new scenarios.
Research, innovation, application…
Frequently, technological advances follow a pattern in which initial development is done by highly specialized research teams part of corporate innovation departments, universities and public or privately funded institutions (Google Brain, Stanford NPL Group, TTI Institute…). Knowledge generated by research teams is initially shared in specialized forums and remains out of the focus of the general media for months or years. Subsequently, the research that comes to fruition produces innovations that are then passed to a wider public, although it remains specialized. Finally, the innovation is transferred to applications and services widely available to the public.
In the case of Generative AI, the first wave of applications is already available and we are beginning to see their capabilities, some of them are:
- Chatbots and virtual assistants powered by Generative AI are already being applied to improve the customer service experience in many companies.
- Machine translation powered by LLMs has greatly improved the way of communicating and the possibilities.
- Media outlets are already using these models to create headlines, summaries and automatically generated content.
- Processing a large amount of data, offering useful information and answers through text, images, videos, and other formats, allowing consultation about them (including unstructured data such as documents) through conversation interfaces and summaries.
- Help with repetitive tasks (responding to requests, locating content, fulfilling contracts, etc.)
Necessary debate
Like other new developments, Generative AI modifies the landscape of possibilities in many individual and business activities, prompting a reevaluation of the balance between possibilities and challenges in emerging applications. But, in contrast with previous innovations, the wide scope of impacted activities and the short time frame on which the changes will take place makes it especially important to pay attention to how our lives may be affected.
Many topics are being debated regarding potential side effects of AI. The question of truthfulness is one of the first to fuel the debate; the connection of automated processes affecting our daily lives with digital channels and AI algorithms raises additional ethical issues like the impact of bias, discrimination, delusions and, ultimately, human participation in decision-making that impacts people’s futures. In addition, the collection and use of personal data during training and subsequent model deployments raises concerns about the privacy of individuals. Or security or intellectual property issues.
A public discussion also arises about the quality and accuracy of AI-generated content. Disinformation and fake news (already an issue before Generative AI) can be amplified using the new algorithms: the generation of realistic text and images makes it more difficult to distinguish between human-generated and machine-generated content, leading to debate about the authenticity of the information we consume.
Organizations and society as a whole demand to establish mechanisms and forum in which to address these concerns in order to successfully manage opportunities and risks.
Regulation
In an effort to strike a balance between technological innovation and the protection of individuals’ rights, governments are stepping up their efforts to establish regulations in the field of artificial intelligence. An example of this is EU’s AI Act in a wide-scope regulation effort affecting companies that provide services in the European Union, and many other countries are also approving their own national AI laws.
By taking responsibility for beneficial development and implementation, governments are engaged in developing frameworks that address ethical issues, transparency in algorithms, data privacy, and the prevention of bias and discrimination. These regulations seek not only to foster public confidence in artificial intelligence, but also to ensure that its deployment contributes to the well-being and progress of society.
Generative Artificial Intelligence Association (GenAIA)
Generative Artificial Intelligence and LLMs will have an increasing impact on society; they are already doing so at a faster rate than any other technical advance we have known. Early awareness and discussion of their possibilities and challenges is essential to enhance the positive impact in our respective fields.
GENAIA, the Generative Artificial Intelligence Association, has been established as a meeting place for companies and individuals interested in AI innovation. Our purpose is to facilitate interaction, knowledge sharing and constructive debate around the progress of Generative Artificial Intelligence. In this forum, participants collaborate to explore innovations and applications, share best practices that promote ethical and responsible use of the technology, address the impact on employment, relevant regulatory issues and, fundamentally, work towards embedding the benefits of generative AI across society.
In this first post, we would like to welcome the first partners and encourage organizations that share an interest in Generative Artificial Intelligence to join the project and collaborate in our activities.
Author: Manuel Vigil, Official Expert of GENAIA
Co-author: Silvia Tamayo, representative of Stratio as an Official Expert of GENAIA and Editorial Board Leader of GENAIA
