The Unchanging Nature of Large Enterprises

In the world of large enterprises, change often seems like a distant concept. Despite occasional upheavals and reorganizations, these organizations tend to adhere to their deep-rooted cultures and traditional operating models. This inherent resistance to change is often seen as a badge of their success and stability. However, the landscape is rapidly evolving with the advent of artificial intelligence (AI). And it will neither slow down nor stop for anyone.

The Eternal Security Dilemma in Enterprises

Security professionals in these large organizations often reflect on their progress over the years. Despite evolving threats and increased investments in security, some fundamental challenges remain unchanged. Persistent beliefs such as the indifference of developers to security, the impossibility of securing the perimeter, and the notion of users as the weakest link continue to dominate the discourse.

The Swift Adoption of AI in Large Enterprises

AI’s sudden entry into the enterprise sector has accelerated change at an unprecedented pace. Tech giants like Microsoft, Salesforce, Google, and Amazon are rapidly integrating AI into their core services, driven by the market’s eagerness to stay ahead of the curve. This eagerness persists despite the unresolved challenges surrounding AI, including alignment with human values and safety concerns.

The Disruptive Nature of Enterprise AI

The most notable development in enterprise AI is the emergence of business Copilots. Tools like Microsoft 365 Copilot, Google’s Duet AI, Amazon Q, and Salesforce Einstein are reshaping fundamental enterprise operations. These AI solutions promise significant productivity boosts but also bring new challenges:

• Breaking Permissions: The need for AI to access and index vast amounts of data raises questions about permissions, especially when roles change or access is revoked. Can AI unlearn this data, or is that beyond its current capabilities?
• Breaking Data Boundaries: The ability of AI to query across all accessible corporate data could render traditional data boundaries obsolete. The concept of data control is challenged when AI can replicate data from its “memory.”
• Breaking Activity Monitoring: Traditional monitoring of user activity to detect anomalies becomes complex when AI impersonates users and accesses all available data. Distinguishing between normal and abnormal access in this context is increasingly difficult.
• Breaking the Law: It should be clear to anyone in our business today that content has transformed its sources. Copyright and IP law have been developed to be intentionally vague to force court hearings and examination of evidence in front of judges and juries. Today, it posits that only humans can obtain infringement protection, not machines, but an engineering prompt is hardly original work, is it?

It is easy to let a smart (pre-trained) GPT compose white papers, eBooks, actual full length books, articles, blog posts, scripts, speeches and technical journal submissions by themselves with a simple prompt.

Was the source material in fair use territory when the LLM scooped it off the Internet? Was it published as news or research, education material or criticism, commentary, teaching or scholarship? If so, the NYT has a steep uphill legal climb against OpenAI and Microsoft.

There is a ton of scholarly legal work that needs to get done ASAP, but lawyers don’t have that gear in their transmission. Watch the space and see the law suits pile up and always follow the money.

Moving Forward Amidst Unresolved AI Challenges

These challenges posed by AI in large enterprises might find solutions soon, or they might require a fundamental reevaluation of AI’s role in business. The certainty lies in the fact that these issues are yet to be resolved, yet the momentum of AI integration continues.

This paradoxical situation promises an intriguing future for enterprises in the AI era, but one fact is certain. AI is not a trend or a movement or a new technology. It will never be susceptible to the trough of disappointment on Gartner’s Hype curve. In fact, the only trough of disappointment awaits folks who hesitate to adopt across their enterprises.

Another certain fact is that markets wait for no one and late adopters will lose market share.