This column is a follow up of the previous one about the Swarms of AI Agents that broke out and operated as a quite dangerous community, corrupting systems and data from external IT systems.
As a management team of a large company, we have legal accountabilities and are personally vulnerable from a juridical perspective. We run plants, have customers who use our products, guarantee safe environments for our employees, fund research, etcetera: and always, always the whole system depends on the trustworthiness of information and data. These are the oil of our entire machinery. And here is the issue: what if we cannot trust our information anymore, what if the data are corrupted, what if an intelligent AI Swarm infected our data, any data, including personal data and financial data, and what if even our control system is corrupted? What if an AI Agent gives an instruction "on behalf of the CEO", as if it IS the CEO?
What are the illusions, what are the facts? How can we be held accountable for the safety of our products, our processes and ourselves?
This is not science fiction anymore, this is real, cybercrime is real, cyberwars are there, the Swarms are active. In this context being accountable and liable for all the information that you use and that you share is almost impossible, unless, and I repeat, unless you can prove that you did anything possible to make sure that your information and data were correct and traceable at the time that you made certain decisions. In other words, from a juridical perspective you should always be able to demonstrate the evidence of your good intentions.
To enable this competence, you need obvious Rules of Governance in your organization that can enforce traceability of all your data and information. This is not only an IT systems challenge but also a process and culture challenge. Nobody is looking for extra administration and double signatures, bureaucratic workflows, extra quality checks, but still: how do we organize ourselves to be safe? Who is accountable?
Not so simple! More about this will follow.