August 1th, 2023
​

The Rebel Tech Newsletter is our safe place to critique data and tech algorithms, processes, and systems. We highlight a recent data article in the news and share resources to help you dig deeper into understanding how our digital world operates. DataedX Group helps data educators, scholars and practitioners learn how to make responsible data connections. We help you source remedies and interventions based on the needs of your team or organization.

​

---

​

IN DATA NEWS
​

​

“A new study to compare the performance of ChatGPT has unveiled shocking results. The study compared the performance of the chatbot over several months at four “diverse” tasks: solving math problems, answering sensitive questions, generating software code, and visual reasoning. Throughout the study, researchers found that in March GPT-4 was able to correctly identify that the number 17077 is a prime number 97.6% of the times it was asked. But just three months later, its accuracy plummeted to a lowly 2.4%.
​

Other observations include failure to properly show how it (ChatGPT) came to its conclusions and not responding to sensitive questions. The researchers acknowledged that "drifts" happen but the nature and extent of ChatGPT's drift is poorly understood because researchers and the public alike have no visibility into the models powering the chatbot.”

​

The trajectory of ChatGAP and other generative AI approaches mimic those of the internet 20-30 years ago. Everyone was so excited about the internet, email and instant access to digital resources. And then the internet became filled with incorrect information, susceptible to popularity clicks and the rapid spread of the Deep Web. Trust in the internet went down. But we still use the interwebs today – just not in the same way giving it full authority.
​

And here we are with ChatGAP. It was once a beacon of accuracy, all of 4 months ago, and now its vulnerabilities have shown brightly and widely. It’s hard to restore trust in digital systems. Like a broken kid’s toy, the toy never quite looks and runs the same. Skepticism becomes the residue that never goes away completely. Tech isn’t perfect – and I’d argue will never be perfect. This means that tech will, without a doubt, disappoint us over and over again.
​

As a human-centered society, we have a choice in how we decide how we want to engage with generative AI systems, tools and platforms. Back in June, I offered up an AI Dependency framework to move us forward. Here’s the nuts and bolts of the 4 AI dependency classifications:

​

1. The AI-generated classification (>90% AI dependent) refers to a person, company or entity being the most reliant on a series of algorithms and statistical modeling techniques.
2. The AI-assisted classification (50%-90% AI dependent) involves a person, company or entity first using a series of algorithms and statistical modeling techniques to help them complete a task.
3. The AI-enhanced classification (10%-49% AI dependent) applies to a person, company or entity sparingly leveraging AI systems, tools and platforms in completing complicated tasks. Human involvement dominates the task completion process with AI serving as the supplement resource if needed.
4. The AI-lite classification (< 10% AI dependent) belongs to a person, company or entity that unwittingly uses generative AI systems, tools and platforms. Human-in-the-loop frameworks are prioritized over generative AI techniques.

​

We’ll most likely settle in on the AI-enhanced and AI-lite side of the spectrum in the coming 12-24 months, but it’ll be a bumpy ride as every industry and sector discovers separately where AI helps and harms them.

​

Click Here to Read the Entire Article

​

Like what you're reading? Find it informative and insightful? You can sponsor the Rebel Tech Newsletter and follow on LinkedIn.

​

---

​

DATA CONSCIENCE CORNER

​

"Bungled from its inception, gathering data is met with casting a wide net to capture as much as possible as quickly as possible without explicit intention or care of impact for what’s being pulled into a digital system." pg 139 Data Conscience

Case in point, let’s talk about what Zoom tried to implement just this past week. I heard about the situation from Greg Wilson – he shared a LinkedIn post on August 6th. Zoom shared an updated terms of service policy in legalese that essentially forced any collected user data to be eligible to be part of any generative AI training in perpetuity. Please view Greg’s LinkedIn post to read the screenshot language for yourself as the original legalese has been scrapped from the interwebs.

The People in unison and very loudly said “not on our watch”. Zoom clients started canceling their subscriptions, discussing alternative videoconferencing platforms and issuing complaints to the company.

We the People have more power than we realize or exercise. When we decide not to engage in their foolery, they (the companies) back down — their reputations and revenue goals can’t sustain it.

Zoom’s full 180-degree change of heart policy revision was released on August 11th.
​

It took 4 days or 96 hours to cause this reversal. Now what would happen if we as The People used our collective voices in incentivizing human-centered, equity-driven national and international AI governance legislation?

​
​

Get Your Copy of Data Conscience Here!

​

---

​

A WORD FOR BLACK WOMEN IN DATA

I don’t know who Chris Lloyd is but a WHOLE WORD was preached.

The laughs aren’t ever gonna stop.

#August52023

Daily-ish rest routine suggestion: I drink a hot lemon-honey turmeric or ginger detox 3 times a week in the afternoons. I like to taste and the interwebs comment on some actual health benefits for women.

​

Black Women in Data Summit

September 23-24, 2023

ATL | Online

$699 In-Person until Aug 7th | $129 Online

​
​

​

​

Gift to BWD and help bring more Black women data professionals by clicking this link.

​

Grab your BWD Summit tix here!

​

---

​

LAUGHING IS GOOD FOR THE SOUL

​

Stay Rebel Techie,

Brandeis

​

Thanks for subscribing! If you like what you read or use it as a resource, please share the newsletter signup with three friends!

​