What’s Cooking with Generative AI – Three Highlights from My Follow-Up List

Generative AI continues to facilitate the forward march of impressive innovations, transforming how we interact with technology and digital content. As we advance further into 2024, here are three notable developments in the realm of generative AI that deserve our attention:

1. EU Regulation on AI Transparency

With the EU’s historic AI Act receiving approval from member states, the focus is now shifting towards ensuring ethical, transparent, and secure usage of AI. This regulation is particularly significant in terms of protecting basic human rights, while promoting technological innovation.

A key aspect of the regulation concerning the “limited risk” use of AI is that AI systems, such as chatbots, must clearly identify themselves as non-human when interacting with users. This ensures that any information disseminated by AI, particularly on topics of public interest, is transparently labeled as AI-generated. This will require extra vigilance from ie. government officials planning on boosting their chatbots with AI.

Image: European Commission

2. Meta’s love affair with AI

Meta is enhancing integration with AI across its platforms, introducing a suite of AI tools, including image generation tools, AI chat functions, and virtual AI companions. These are currently only available in the US, and it is yet undetermined when or if they enter the EU.

Image: Meta

Starting from May 2024, Meta will begin labeling images and videos produced by AI, with a transition period for compliance. This development reflects a growing trend among tech giants to incorporate AI in ways that aim to be both innovative and responsible, acknowledging the potential risks and ethical considerations associated with AI-generated content.

3. Sound and Fury

The generative AI landscape is also buzzing with developments in voice and image synthesis, including technologies that turn text into video and vice versa. Companies like OpenAI with Sora and Google with Lumiere are pushing the boundaries with tools capable of enhancing video content or generating entirely new visuals. Interestingly enough, Sora is being trained with Shutterstock’s video content – with permission. This is a step in the right direction when it comes to copyright issues at the heart of AI training.

However, the rise of deepfake technology and its implications in terms of misinformation and privacy concerns continue to provoke heated discussion. Tools like Intel’s FakeCatcher are emerging in response, aiming to detect synthetic content and mitigate the risks associated with deepfakes.

Video: YouTube/OpenAI (Sora)

Whether it’s the way we consume information, interact with digital content, or even perceive reality, AI is the nucleus of a significant evolutionary step. As we continue to explore these innovations, it’s crucial to balance enthusiasm for technology’s capabilities with a cautious approach to its ethical, social, and legal implications. Keeping an eye on these developments ensures that we’re not just passive observers but active participants in shaping the future of generative AI.

Kansikuva: DallE