Big Data & Analytics - Thinks & Links | August 12, 2023
BD&A - Thinks and Links
Big Data & Analytics - Thinks and Links | News and insights at the intersection of cybersecurity, data, and AI
Happy Weekend!
Data and security will be the key to winning in this Incredible AI age.
The history of technology advancement is one of ever-increasing productivity at a lower cost with altered risks.
When factories began to allow for the mass produce goods it created disruption, riots, and concerns. The new technology also introduced new risks and concerns for businesses that relied on these machines – that they could keep working. Threats to continued production included having the technical expertise to fix a broken component and also to protect against “insider threats.”
For example, English textile workers did not care for the “fraudulent and deceitful” machines that took away many skilled labor jobs. They gathered to write threatening letters and hold raids to destroy these threatening machines. They named their movement after a legendary weaver “Ned Ludd” and these “Luddites” were a threat to mill and factory owners for some time until legal and military forces intervened. The name lives on as a term for those generally opposed to new technology.
Despite the very real social upheaval from these advancements, they also provided the access for more people to gain access to textile goods and tailoring services. Over time it has become cheaper and easier to get clothing and other factory-made materials.
The same goes for computing. From the earliest room-sized computers to today’s iPhones, the capability and power of technology is getting greater and at lower cost. AI represents the latest wave of this trend.
It has been possible to build generative AI chatbots for quite a few years now. OpenAI’s first GPT model came out five years ago. With the right knowledge and tools you could have built a ChatGPT product a year or two ago that would have performed decently well. But today of course, that capability is much better and completely free. There are Natural Language Processing tasks that would cost thousands of dollars a few years ago that can be completed now for free thanks to the latest LLM models. Companies, universities, and open source researchers are pushing the limits of these tools daily – and doing it at a phenomenally cheaper cost.
Play this scenario out a few years into the future, and there are two things that I believe will be fundamental to success with this technology:
Perhaps my theory of AI economic development aligns best to Syndrome the supervillain in the movie “The Incredibles.” Recall, his evil plot focused on using his company to develop advanced weapons and robots that could allow world governments to have more powerful capabilities than the super hero’s. He captures our heroes and monologues (as all supervillains and “thought leaders” must) about how “when everyone’s super… (evil laugh) no one will be”
Of course, (spoiler alert), the heroes eventually win and defeat our supervillain. They do this by bringing their human element to the fight as a family who work together. They also have experience (data) fighting super villains and use Syndrom’s own technology against him (security).
When companies are competing using ever more AI in their technology stack, the ones who will prevail will do so by leveraging the unique data and experience that they have serving customers. They will also need to avoid self-sabatoge and the increased risk that a changing technology landscape will provide.
Data and security will be the key to winning in this Incredible AI age.
Wall Street Journal’s CIO Journal: AI Is Generating Security Risks Faster Than Companies Can Keep Up
In this article the Wall Street Journal summarizes much of what we cover in Optiv’s AI Executive Briefing. Technology leaders are recognizing that demand for generative AI capabilities are substantial, and the use of these models introduce risks that must be addressed before going into production. The article includes:
There are solutions for these concerns and many organizations are building the governance and controls necessary to mitigate the risks.
Zoom Had a Bad Week
https://www.axios.com/2023/08/09/zooms-terms-service-changes-ai-fears
Zoom recently changed their terms of service to grant the company extensive rights to use content from meetings to train its AI algorithms. This small edit created a big uproar. The company later put out a blog post, stating that they would not use audio, video, or chat content for AI training without customer consent. However, there are still important questions involved with how granting and revoking consent would work.
As AI becomes easier for more businesses to access, the way they use data can be a strategic advantage. But a poorly constructed set of policies and communications or security concerns can torpedo a business’s AI aspirations.
Azure OpenAI Service Achieves FedRAMP High Authorization
https://devblogs.microsoft.com/azuregov/azure-openai-service-achieves-fedramp-high-authorization/
Content creation, summarization, and advanced automation capabilities are coming to Federal Government customers now that Microsoft has certified their Azure OpenAI service as FedRAMP High. This classification is the standard for security necessary to protect some of the federal government’s most sensitive unclassified data in cloud computing environments. It includes clearance to be used by law enforcement, emergency services, and healthcare. It is essentially stating that the government trusts this system to be operational and secure even for services that could impact economic activity or pose threat to human life.
This is a significant milestone for trusted generative AI services which will see more and more public sector uses of the technology while also assuring private sector CISOs that secure solutions are possible with the right framework in place.
Unlocking Real EBITDA Value with Generative AI
https://insights.alixpartners.com//post/102ik6t/unlocking-real-ebitda-value-with-generative-ai
We focus a lot on Risk in this newsletter, but there’s an important but un-categorized Risk of doing all this work with Data and AI without focusing first on the value. Investors expect business operators to create more value using technology, and if we are going to start prioritizing which things to build and secure it helps to have a good framework on value. This article nicely summarizes the steps to do that:
-
Answer foundational questions:
-
Then look at which EBIDTA enhancement lever is most relevant to your business:
Have a Great Weekend!