This Series of AI Models Are Built on Human Contractors – The Next Will Be Trained On Computer-Made Data

In a world being rapidly digitized, Artificial Intelligence (AI) has quickly emerged as a frontrunner in revolutionizing how we interact with technology, but the tech industry could be on the verge of a revolution in how AI models are developed.

The crux of this transformation lies in how AI is trained, today, AI technology relies heavily on human-compiled data sets to function effectively, a method that is rapidly evolving.

Let’s delve into the intriguing transition from human-built datasets to computer-generated ones.

The Current Landscape: Human Contractors Behind AI

In the realm of AI, the limelight is often cast on high-profile developers and innovators.

But, the foot soldiers working tirelessly in the background – the human contractors – are the unsung heroes who make AI functionality possible.

Their role is to scrutinize AI responses meticulously, amending inaccuracies, and eliminating biases to provide users with credible and reliable information.

These individuals, however, are grappling with significant challenges – from paltry wages and excessive workloads to stressful conditions and fear of job security, the human element in AI development is under tremendous strain.

This strain not only impacts the well-being of these contractors but may potentially affect the quality of AI user experiences, and yet, their contributions are indispensable, enabling AI chatbots to function seamlessly.

ChatGPT, an AI model developed by OpenAI, exemplifies this human-machine synergy – the impressively interactive and adept responses this chatbot delivers is a product of tireless contractors working in the backdrop, improving the models’ responses, hour after hour.

LLM datasets be like:
• First you start with CommonCrawl
• Then you add C4, which is just CommonCrawl again, but dont worry about it
• Eye of new̸̮̺̿͝ͅt̷͚̆͊́̀t and Toe of f̶̝̈͊̊̚r̸̦͕̻͇͒̕͝ỏ̵̝̳͖̘́̃͝g
• Then Wikipedia (also is in CommonCrawl but dw)
• GitHub (permissively…

— swyx @ (@swyx) May 1, 2023

The Paradigm Shift: Towards Computer-Made Data

Yet, the AI industry is not one to remain static, and the next generation of AI is already taking shape.

This shift signifies a move from human-built datasets to computer-generated ones, and while this may appear to side-line human contractors, it opens a new frontier of possibilities.

Using computer-generated data will provide vast volumes of information for AI to train on.

This change could result in more robust AI models capable of learning and adapting more swiftly, autonomously, and efficiently than ever before.

Moreover, this transition could mitigate the challenges faced by human contractors, making AI development a more sustainable and scalable process.

Ethical Implications of Transition to Computer-Made AI

Speaking to the FT, AI pioneer, Aidan Gomez, believes the future of AI will manifest in its widespread and ubiquitous use, becoming as fundamental to our lives as our smartphones.

He envisions AI serving as an invaluable assistant that simplifies complex tasks, streamlining online experiences, and transforming industries like healthcare by reducing human errors.

Nevertheless, Gomez insists that the role of humans in the AI world will remain vital, particularly in high-stakes sectors such a medicine.

For now, AI will work to assist humans, not replace them, ensuring an added layer of assurance for critical tasks and decisions.

But as we advance into this new era of AI, the potential for innovation as well as misuse is vast.

This highlights the importance of ensuring ethical practices and risk mitigation strategies are in place to prevent misuse, an approach already being mainlined by the EU.

As Gomez states, it’s about integrating the new technology into our lives responsibly and conscientiously – and this could be a big topic at an upcoming summit on global AI regulations in Autumn.

The transition to computer-made data is indeed a significant leap for AI, but it’s a leap that must be taken with caution and responsibility.

While the journey continues, and the next chapter of AI evolution is about to unfold – stay tuned with B2C for the latest AI news.

RELATED:

• Vatican’s AI Ethics Book: Guiding the AI Revolution
• Elon Musk Says His AI Startup x.AI Will Be A Completely Unbiased Champion of Transparency – Is This a Good Idea?
• OpenAI Jobs: Can You Steer Superintelligent AI to Follow Human Intent?