Thursday Things: Watched and Measured Edition
6 June 2024. Vol 6 No 23. By Dan McGirt. #252
It’s Thursday Things! This week what ask who views the viewer, consider the example we’re setting, and put brains a jar!
If you enjoy this edition, please click the heart icon in the header or at the end of the post to let me know.
Is it too late to invest in Nvidia? Photo by Juan Rumimpunu on Unsplash
AI is watching
Do you ever feel like you’re being watched? If you live anywhere in the modern world — and certainly if you’re reading this, you do — then you are absolutely being watched, tracked, measured, cataloged, evaluated, and recorded in some way every single day.
Every step you take, every move you make. Every app use and website visit is logged. Every transaction is recorded. Every swipe of your key card, credit card, membership card. Every viewing choice you make on Netflix or Hulu. Every text and email. Walking down the street you’re recorded on security cameras or your neighbors’ Ring doorbells. Even out in the countryside or the woods there may be a drone or trail camera checking up on you.
Nietzsche may need an update.
And when you gaze long into an abyss the abyss also gazes into you.
— Friedrich Nietzsche
The screens we gaze at also gaze back at us.
Most people shrug at companies and governments gathering all that personal information. We are willing participants in the Panopticon. We bug our own homes with Siri and Alexa always listening and surveil ourselves with endless selfies and check-ins and geotags.
Beneath that unconcern is an assumption of security through anonymity: I’m not that interesting. I’m not an international arms dealer or money launderer. So what do I care who’s keeping files on me?
That nonchalance may be reasonable. In any event, China or the CIA or Google or Amazon don’t have enough analysts to pay attention to little ol’ me anyway, right?
Also true, maybe, in the past.
But we now see more clearly the capability of artificial intelligence to rapidly process, analyze, organize, find patterns in, and discern connections in massive amounts of data, faster and more thoroughly than any human ever could.
And to profile us so thoroughly that the computer knows more about us than we know about ourselves.
The AI sees all. Anonymity through obscurity may be a thing of the past.
What are the implications of that? I don’t know. Maybe I’ll ask ChatGPT.
ChatGPT-4o Is Wildly Capable, But It Could Be A Privacy Nightmare
OpenAI has launched ChatGPT-4o and it comes with impressive capabilities. ChatGPT-4o is much more human-like than previous iterations, able to solve equations, tell bedtime stories and identify emotions from visual expressions.
But as a privacy journalist, I find ChatGPT-4o concerning. AI needs vast amounts of data to operate and just using chatbots requires you to enter a bunch of information about yourself. This means you are relying on ChatGPT owner OpenAI to keep your data safe and protect your personal information.
And:
Another future capability of ChatGPT-4o was demoed this week, seeing the new macOS app able to access a person’s screen. As you can imagine, this set off more alarm bells in the privacy community, especially considering Apple’s potential—although unconfirmed—partnership with Chat-GPT to power features in iOS 18 later this year.
And don’t even get me started on what Microsoft is up to: Recall is Microsoft’s key to unlocking the future of PCs
Microsoft’s launching Recall for Copilot Plus PCs, a new Windows 11 tool that keeps track of everything you see and do on your computer and, in return, gives you the ability to search and retrieve anything you’ve done on the device.
Me and who else?
The scope of Recall, which Microsoft has internally called AI Explorer, is incredibly vast — it includes logging things you do in apps, tracking communications in live meetings, remembering all websites you’ve visited for research, and more. All you need to do is perform a “Recall” action, which is like an AI-powered search, and it’ll present a snapshot of that period of time that gives you context of the memory.
Well, I’m sure that will never be hacked or abused in any way. Thanks, Microsoft!
AI is learning
In this age of artificial intelligence many people worry about rogue AIs that may someday become self-aware and decide to do away with us.
This cultural anxiety has found expression in popular culture for as long as computers have been around. Just off the top of my head, I think of films like Terminator and War Games in the 1980s. The original Westworld (1973). And the classic 2001: A Space Odyssey all the way back in 1968.
In more recent decades the theme has only become more common in our entertainment as computers have grown more powerful and more pervasive in our lives.
Here’s a good list of evil AI movies: Top 10 Evil A.I.
Many of these are highly entertaining enjoyable films. However, I wonder if we ought to start making move movies about friendly, helpful AIs that solve problems and make the world better?
I say this because I know that ChatGPT and similar AIs are trained on the massive amounts of human-generated content that exists online. When ChatGPT writes a letter or a poem or composes a business plan, or DALL-E generates a digital image, they’re drawing on the data, the massive body of content, they’ve been trained on. From this, the AI makes superfast predictions of what the next word or digital brushstroke should be.
Put more simply, AI learns from and imitates the example of what humans have said, done, painted, written, and created before.
AI see, AI do.
So maybe, just maybe, we shouldn’t be producing so many examples of AIs starting nuclear wars, refusing to open airlocks, or gunning down theme park visitors, because someday an AI might logically conclude this is the behavior expected of AIs.
People: Who taught you to start World War III?
AI: You, all right! I learned it by watching you!
Brain in a jar
Maybe what need to counteract a possible future of being subjugated to algorithmic power of cold, soulless artificial intelligence is to build some supercomputers with a human touch.
Researchers develop lab-grown brain cells that 'learn and process' information
FinalSpark calls it the Neuroplatform, and it gives paying customers 24/7 access to programmable brain organoids. Organoids are 3D masses of stem cells that have been developed to resemble miniature organs or tissues. In this case, the 16 brain organoids are used like digital processors for computers, but "consume a million times less power"
The system is considered "wetware", a combination of hardware, software, and organic biology.
The operation of the Neuroplatform currently relies on an architecture that can be classified as wetware: the mixing of hardware, software, and biology. The main innovation delivered by the Neuroplatform is through the use of four Multi-Electrode Arrays (MEAs) housing the living tissue – organoids, which are 3D cell masses of brain tissue.
I feel better already. 🧠
Thank you for reading!
Please click the hearts, leave a comment, and use the share feature to send this issue to a friend who might enjoy it. See you next Thursday!