A Novice's Journey into Artificial Intelligence

Author: Yaw

“This is AI at it’s worst”

This rap song is a back-and-forth featuring two voices. One voice is a guy called AllttA. The other voice is one that sounds like Jay-Z. But, the “Jay-Z” is not actually the rapper Shawn Carter, but an AI generated voice.

For a guy who has consumed as much hip-hop as anyone else, this is incredible.

Continue reading

Dope Pope

As festooned with jewelry and precious metals the Popes of the past were, I doubt you’ve seen one this swaggy…

When you’re blessed by the Lord. Source: Reddit.

And the reason why is because it isn’t real. This image is created via an AI engine called Midjourney. You can find coverage of it all over the internet: The Verge, BuzzFeed, Reuters, and ArsTechnica to name a few. And if you read them you’ll find that they address a bevy of concerns relating to generative AI.

In this image, I see not only a really cool Pope, but also the promise and the peril of artificial intelligence.

What’s the promise? It’s sort of fun!

Pope Francis in a Balenciaga bubble coat. It’s super unique and refreshing. Here’s someone who is the literal manifestation of an institution that many consider to be stiff, stuffy, unchanging, patriarchal, hierarchical, and anachronistic. And he is wearing clothing that is the entire opposite of that.

It reminds me of the novelty of the early internet. Way back when you could find a cute viral video and accept it at face value. You wouldn’t immediately wonder if it was staged. Or if it was ripped off another website. Or if it wasn’t simply monetized click-bait. We didn’t know it, but the internet was fun, innocent, and enjoyable without ulterior motive.

Made in God’s image. Source: Reddit.

What’s the peril? Well… how much time do you have?

Let’s admit that differentiating betwen truth and falsehood has always been a large concern for human kind. I would argue that this represents a dramatic shift in what we readily accept as true. When I first saw this image on Twitter, I immediately thought, “damn, this guy looks cool.” Then I immediately thought, “no way this is real.” But I really wasn’t sure! Deep fakes have been around for a while. But this is a solid articulation of how deceiving they can be.

While this is cool, can you imagine how this might be used in an irresponsible, harmful, or evil way? It wouldn’t take much to imagine His Holiness being falsely presented in a problematic or compromising situation.

How do you build guardrails to prevent or mitigate this type of image? I don’t have insight on this, but I understand that OpenAI and others place specific restrictions on improper use of AI. But will others follow suit?

How do we fact-check this type of image? There are some telltale markers of AI generation of this image — the smears, the slightly hidden right hand, and the random coffee mug he’s holding. But without specifically knowing that signs, it’s easy to accept this as fact. Moreover, we know that the AI engines will continue to improve and “fix” these “errors.””

Also, how do we address the use of one’s likeness? Does the Pope have a right to not have his image used like this? Do you and I as private individuals have to provide consent to have our image run through an AI neural network? Do we only have rights when our image is easily discernible in the end result? Do our rights only extend to immoral or abusive use of our likeness?

His Swagginess. Source: Reddit.

Seeing is believing, right? But what can we believe when this technology becomes commonly available? How easy will disinformation be to create? Will there be liability for deceit and lying? So as much as I enjoy the prospect of seeing Pope this dope, I also recognize (like many others) that there’s a lot we need to do to prepare for this new technological wave.

Great Vergecast podcast episode on generative AI

If you don’t already listen to The Vergecast podcast, I highly recommend it. They cover the broad intersection of technology and culture in the most engaging way possible. Recently, they had an episode where they discussed the ins and outs of generative artificial intelligence.

What is generative artificial intelligence (generative AI)? In simple terms, it’s a tool that uses artificial intelligence to create content that matches (or comes close to matching) what a human would produce. I’ve typically seen machine learning deployed in this way, but the end result can be an essay, a corporate slogan, a fantasy-like painting, or even a very realistic looking photograph.

Imagine – you want a stock photograph of a professional meeting. Normally, you’d hire a professional photographer and paid actors to capture the scene of well-dressed business types sitting, smiling, and talking around a conference table. With generative artificial intelligence, you can simply type in the scene you want captured, and the AI produces it for you.

The Vergecast episode was great in that it provided a few glimpses into some critical issues about generative AI. They get into what the various AI models are, who is actually creating these tools and models, how the models are trained, what is produced, and the ethics and economics of all the above. There are some really important questions that they highlight, like:

  • the veracity of content,
  • ownership of training data and the potential infringements on those ownership rights (i.e. copyright infringement),
  • liability for making models available versus liability for using them, and
  • protections and guardrails around these tools.

It’s all interesting stuff, and there’s so much more that needs to be sorted out over the coming years.

Additional reading:

Intro

Hello world.

I’m not sure how long this blog will last. I’m not sure of how good it will be. I’m not sure how frequently it will be updated. All I am sure of is that I am interested in artificial intelligence (AI), and I don’t know much about AI. Thus the name of this blog: IDKAI, or “I don’t know artificial intelligence.”

I’m a novice in the AI space. My exposure has been limited to reading a couple of books, seeing the news headlines and articles, and using some of the consumer-facing virtual assistants (like Apple’s Siri or Amazon’s Alexa). I want this blog to be a vessel to encourage my interest in AI, document some of my views and thoughts on the subject, and maybe share the world of AI with others (whether they’re fellow novices or passing observers).

So if that interests you at all, feel free to stop by every once in a while, and share your thoughts. I hope you enjoy.

“An Arab scholar” by Rudolf Ernst (Artvee.com)

© 2024 IDKAI

Theme by Anders NorenUp ↑