How generative AI can hurt cloud operations

Generative AI can generate new content, and it’s been heralded as a groundbreaking technology with the potential to transform various industries. However, those working in the cloudops world who will be charged with running generative AI systems long term are beginning to voice their concerns.

Although generative AI has many benefits, it also has the potential to cause harm to cloud computing operations. Today these are theoretical problems, but they will soon become a reality. Thus, it’s helpful to talk about some of the more concerning issues before we fall in love with this technology—or at least prepare to tackle some of these issues before they cause real problems.

Security risks

Generative AI can be used to generate fake data that can fool cloud computing systems. This fake data can launch attacks on the system or manipulate the system’s behavior, leading to security breaches, data leaks, and other security risks. Additionally, generative AI can create fake identities that can circumvent security measures and gain access to sensitive data.

Powerful tools can do as much harm as good. Generative AI is no exception. I expect to see many future breaches driven by generative AI. New and more expensive AI-powered cloud security tools will combat these breaches. See how this works?

The value you gain from generative AI can be quickly outpaced by the increased security requirements to contain generative AI interference from outside sources. An enterprise that realizes no gains from the internal use of generative AI will still have to pay to protect itself from generative AI-powered attacks on its security systems.

Resource overutilization

Generative AI algorithms can consume significant resources, leading to the overutilization of cloud computing resources. We’ve already covered this issue. You might see slower system performance, reduced system availability, increased costs, and more carbon produced. If generative AI algorithms are not optimized for cloud computing environments, they can cause a significant strain on the systems. It will fall on the cloudops staff to fix the resulting problems.

Incompatibility with existing systems

Generative AI algorithms can be incompatible with existing cloud computing systems, leading to integration issues. This can delay the deployment of generative AI algorithms and cause problems with system performance or efficiency.

I have significant concerns about this, but I’ve not seen the same level of unease from people deploying generative AI systems who must integrate intercloud and intracloud systems. I suspect this will emerge as one of the more complicated operational issues, as integration is usually the sticky wicket.

Unpredictable behavior

Generative AI algorithms can exhibit unpredictable behavior, which leads to unexpected outcomes. This can result in system errors, degraded system performance, and other issues that are impossible to predict. I suspect we’ll get better at predicting behavior as we learn more about generative AI system operations, but the learning curve will be painful. I’ve already had some generative AI systems pulled off cloud systems due to unpredictable behavior and, what’s worse, unpredictable cloud computing bills.

Generative AI is an unstoppable force in the enterprise technology space. It’s yet another technology made more accessible and affordable by cloud computing, and the easy availability of this technology will reverberate through the marketplace. Generative AI will become a technology that allows businesses to succeed by out-innovating their competition.

Although generative AI has many benefits, it also has the potential to create many problems for the cloudops team and automated systems. As generative AI continues to be developed and deployed, it is essential to consider these potential risks and take steps to mitigate them. I suspect that few developers are considering the drawbacks at this point. Trust me, the impact of this technology will soon be felt in good ways and bad.