Note to CISOs: How to stop ‘shadow’ AI>
SC Media –
The emergence of generative AI has sparked both fear and optimism among experts and the public.
Some concerns include the potential for erroneous results that could harm specific groups or spread disinformation, the fear of job automation, and the worry that generative AI could lead to job losses and even the downfall of humanity.
In the enterprise sector, C-suites are worried about the exposure of private data to public large language models (LLMs), which could compromise sensitive information.
Companies like Samsung have already experienced data leaks and have implemented security policies to control the usage of generative AI tools.
The decision to reject AI adoption can lead to shadow AI, where employees find ways to use AI tools without proper controls, posing security risks.
History has shown that users will seek ways to leverage new technology, even without official approval.
It is important to give employees access to generative AI innovations to democratize cybersecurity, enhance productivity, and personalize user experiences.
However, strong principles and careful selection of LLM models are necessary to ensure effective and secure deployment.
When implementing AI, organizations should prioritize the security and protection of customer data and consider the accuracy, reliability, trustworthiness, and bias of the data that powers the LLM.
Controlling access to generative AI capabilities is crucial, with features such as transparency, exclusion and anonymization of data, and auditing and compliance controls.
Collaboration between humans and AI technology is key to addressing fears about job replacement.
By embracing AI adoption, safeguarding data, using reliable data sources, and implementing necessary controls, organizations can minimize risks while empowering their workforce with the benefits of generative AI.
Overall, a continuous focus on AI data privacy and quality will help society become as comfortable with AI as it is with cloud computing.
Link: https://www.scmagazine.com/perspective/note-to-cisos-how-to-stop-shadow-ai