Introduction to Generative AI by Numa Dhamani and Maggie Engler.
Numa holds degrees in Physics and Chemistry from the University of Texas at Austin. She has served as the Principal Investigator on the United States Department of Defense’s research program. Today she is a Principal Machine Learning Engineer at Kungfu.ai.
Numa is an adjunct instructor at Georgetown University.
Maggie holds Masters in Electrical Engineering from Stanford University. She is a safety team member at Inflection AI. Previously Maggie was a Data Science Fellow at the Center for New Data and a Senior Machine Learning Engineer at Twitter. Maggie is also an adjunct instructor at the University of Texas at Austin School of Information.
Perhaps this book should be required reading for any organization’s AI planning team. Yes, your AI staff should have this well understood. However Numa and Maggie convey a grounded understanding of large language models (LLMs). In addition, you will understand the new rush for integrating generative AI (Microsoft and OpenAI) into your organizational workflows. To be fair, they also are addressing organizational benefits, risks, and limitations.
LLMs can go sideways fast
They are certainly providing a solid ground on LLMs and training models. Both the impressive capabilities that have taken the world by storm and the vulnerabilities are front and center. They are addressing vulnerabilities from training data. Thankfully Numa and Maggie address data privacy considerations and regulations regardless of your organization’s marketplace. The impact from data leaks to European data policies help shed light on how organizations must carefully consider the best vendor to engage and not simply rush to jump onboard the ChatGPT train. Just ask Samsung.
The impact across Education will be far and wide
I was drawn to ‘Generative AI’s footprint on education’ and how they address OpenAI’s classifier GPTZero. At the time of this writing the focus on ‘traditional’ chatbots did not include the more recent announcements of AI Agents. However, to be clear it is rather unnerving to see educators in both K12 and HigherEd jumping on the social media influencers bandwagon. There are wider issues to address within state education systems including the impact of synthetic content. Numa and Maggie certainly deliver good insights.
Criminals will always be two steps ahead
Their credibility score is up as Numa and Maggie have a chapter addressing adversarial attacks. This is the world we engage today. Not addressing vulnerabilities would lessen this book’s impact. I am glad they provide an overview to adversarial attacks, the misuse of chatbots, and AI hallucinations.
Since the release of ChatGPT 3 in late 2022, our globally connected world is observing yet another rapid innovation. And the race to capitalize this marketspace is very competitive. However, this is without understanding the impact and consequences. It is no longer enough to rollout a promising technology service and not expect criminals and nation states to quickly find vulnerabilities.
We are certainly just scratching the surface. I believe readers can benefit from The New Automation Mindset. This would seem to be a good follow up book.
In conclusion, Numa and Maggie are providing a much needed and well understood introduction to GenAI. This is a well written book and worth your time.