October 24th, 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 in understand 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

​

​

“Security researchers have discovered a new type of attack against LLMs called "indirect prompt injection.” This attack allows hackers to manipulate LLMs by concealing instructions in web pages or documents, causing the LLMs to behave in unintended ways. The cybersecurity industry is concerned about the potential risks of these attacks, as LLMs are increasingly being used by large corporations and startups.”​
​

Lovely, yet another tech-created AI problem that they are ultra concerned about the risks it poses. You know, this prompt injection issue sounds eerily familiar. Oh what could it be? Hmmmmmm, let me think…

SQL injection.

SQL injection is a common attack used by bad actors to query a database management system (DBMS) with bad SQL statements. These SQL statements are designed to inadvertently reveal the structure and/or contents that exist in the database. For instance, you can create an SQL query that could reveal the first and last names and the encrypted passwords of all people with admin in their job titles. These encrypted passwords could be decrypted using another software, which would enable the bad actor to obtain a copy of all the data accessible to those senior leaders of that organization. Ha, now you know the basics of how a data breach happens!

So how are SQL injection and prompt injection different/similar? Mostly English-based text, instead of SQL code, is used by bad actors in prompt injection. That’s it. SQL injection is a very well-known problem in the DBMS world with a slate of impacts and prevention methods. The impacts include but aren’t limited to: sensitive data exposure, organization’s reputation decline, data loss and denial of service. The prevention methods cut across DBMS development, question answering and production phases.

It’s a bit baffling why prompt injection researchers aren’t identifying how they could recycle and adapt approaches that minimize the harmful effects of SQL injection and apply it to prompt injection. Doesn’t that involve google searches and watching youtube videos?

​

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

"No matter how we try to reframe, restructure, or reimagine data work, harms persist through the metrics and analytics options." pg 164 Data Conscience Corner
​​
Silicon Valley has lost control of the ‘AI is benevolent’ narrative due to data privacy concerns that necessities AI regulation legislation. Simply put, the AI conversations turn into questions about where the data input to these AI systems originated and whether these AI systems have the ethical/legal rights to that data. So discussing AI is talking about using people’s data unauthorized. The U.S. National Institute of Standards and Technology (NIST) has created an AI Risk Management Framework. It’s a step to define the social and technical tensions generated by AI systems, tools and platforms. But it falls short of suggesting approaches to minimize or avoid introducing harms to people, organizations and industries. Here’s one concrete ethics step shared in Data Conscience that every data/software can do, starting today: collect, evaluate and update the team’s tech stack in order to remove algorithmic-based tools that cause ethical issues, every 12-18 months.

​

Get Your Copy of Data Conscience Here!

​

---

​

A WORD FOR BLACK WOMEN IN DATA

​

It’s never too late to ADORE you. It’s never too late to BE you. It’s never too late to COMFORT you. It’s never too late to DELIVER you. It’s never too late to ENCOURAGE you.

This is the data you need to tell yourself every day until 2024. Your data skills, expertise and lived experience is over-enough to secure that position, negotiate that promotion and/or attract paying clients.

Daily-ish rest routine suggestion: I drink 2 8-oz glasses of warm water with 2 teaspoons of lime juice when I first wake up. It helps me anchor my day — ushering it in peacefully.

​

​

No matter your career goals, the answer is likely the same: why, YES YOU CAN! The question is how do you effectively ask for what you want. Leondra R. Gonzalez is bringing her nearly 10 years of data scene experience in weaving amongst roles as a practitioner, researcher and educator to BWD as our inaugural BWD Community Seminar speaker. She’ll be helping us leverage the power of ask to get ahead, including asking for a raise, negotiating pay, asking for time off, inclusion in projects, and so much more.
​
You’ll be able to ask your burning questions prior to the live seminar and Leondra will address the top requests. tips on how you can advocate effectively for yourself and do it confidently.
​
Join us live or catch the replay whenever you need a healthy reminder of the power of your words. All are welcome to join us! Connect with Leondra R. Gonzalez on LinkedIn.

November 6, 2023

11:00AM - 12:00PM ET.

​

​Register for The Power of Ask

​

---

UPCOMING EVENTS

​

​

OLC Accelerate 2023 | Washington, DC

​

Are you in the DC area? Come out to the Online Learning Consortium Conference happening now! I’ll be delivering a keynote address tomorrow afternoon in-person.

Keynote address: Teaching with Generative AI: Helping Digital Educators and Leaders Better Access Learners’ Knowledge

AI has infiltrated society, growing concerns over its future implications within the education space. AI-assisted tools and platforms, like ChatGPT, are disrupting how instructors are able to assess how much a learner knows and can apply without advanced technological assistance. Whether concerns stem from trying to understand what these tools are capable of or tackling the ethics of students cheating with generative AI assistance, educators and the education community at large (board members, administration and parents) must understand the nuances of these tools.

In this talk, Brandeis Marshall discusses the tensions of contending with the AI-assisted tools’ consequences, institutional use/non-use of these tools, the rapid pace of AI’s new releases, and learner engagement for all levels of education. Brandeis will share small non-disruptive changes for a range of online educators, from those who are wearing multiple hats in the classroom, to those who lead a large teaching staff. Instructional designers, administrators, and digital learning leaders will learn how to navigate the new era of AI-led learning assistance and how to best both navigate and integrate these tools for better learning outcomes.

​

​Join Me at OLC Accelerate 2023​

​

Follow us on social

​

---

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!

​