Artificial Intelligence (AI): Computer-Generated Storm
Artificial Intelligence (AI) has taken the tech world by storm, becoming a central topic of conversation. Once a niche interest, it now concerns everyone. Machine learning and algorithms have long dominated this space, but AI promises something even smarter. These advanced models claim to produce human-like language, art, videos, and voices at impressive speeds. However, human input remains essential for their output. But do these advancements truly represent intelligence?
The Risks of Human-Like Output
The companies behind AI often focus on potential future capabilities, promising remarkable achievements with more data, control, and power. However, this human-like output poses significant risks. For example, scammers now use AI-generated voices to mimic family members, creating new avenues for fraud (The Guardian). These scams involve spoofed phone calls, claiming a family member is in trouble and needs money—an old scam with a new technological twist.
Artists are also under threat from AI-generated cover bands, which dilute their original work (Slate). Similarly, authors face fake books published under their names, damaging their reputations as they struggle to remove these fraudulent works. Deepfake technology exacerbates these issues, escalating concerns around online abuse and privacy (Boston University).
Data: The New Goldmine
Data has become the new goldmine, with AI models demanding more of it. Companies like Reddit have struck deals with Google and OpenAI to license content, setting a precedent for others (Reuters, The Verge). However, some companies are pushing back against their content being used for free to train models. For instance, the New York Times has filed a copyright lawsuit against OpenAI (Harvard Law).
This is just one of many ongoing legal battles. OpenAI’s consumer-focused chatbots and Meta’s open-source Llama model illustrate the rapid evolution of large language models. Open-source models, often developed without much fanfare, are making quick progress. Practical applications like NVIDIA’s DLSS 3, used for upscaling and frame generation, showcase AI’s potential, despite ongoing challenges. Here’s a video showcasing DLSS 3.
DLSS 4 is coming soon that video is two years old and we’re currently on version 3.8.
The Reality of AI Hype
Despite these advancements, much of the current AI hype feels like a marketing gimmick rather than a transformative tool. AI outputs have their uses but are also susceptible to misuse. Concerns about unregulated use by individuals and potential data breaches are already surfacing, as seen in the case of Star Health’s data leak (Life Insurance International). Imagine the risks when employees use such technologies without their employers’ knowledge, potentially exposing sensitive information to remote servers (Forbes). Something that has already happened leaking of information.
Addressing the Challenges
While I don’t see this wave of computer-generated content as an immediate threat, I have serious concerns about our ability to grapple with existing problems. Misinformation and disinformation thrive in an environment that prioritises engagement over quality, exacerbated by lax moderation practices. Tackling these issues would require overhauling business models and restricting who can post online—steps that could reinforce the dominance of existing tech giants and stifle new challengers.
The internet, built on non-profit code and maintained by a few volunteers, faces growing instability. Big tech has built empires atop this fragile foundation, often unwilling to share profits or cede control. With major regulatory powers like the USA, China, and the EU each pursuing their own visions, the once open World Wide Web increasingly resembles a series of closed-off walled gardens.
The Future of AI and Content
With the flood of computer-generated content on the horizon, distinguishing between real and fake will become increasingly difficult. As we silo ourselves into personalised content bubbles, our perspectives on reality may diverge further. Addressing these challenges will require a concerted effort to ensure that AI serves as a tool for enhancement, not exploitation.
For clarity, I did put this into ChatGPT for spell-checking and grammar edits. It is a useful tool but does not transform the world. Here is the problem without me telling you, you would not have guessed.