We’re creating AI that is not secure. It’s not a surprise that we are not securing AI as we find it challenging to secure no-AI products and…

← Back to Blog

We’re creating AI that is not secure. It’s not a surprise that we are not securing AI as we find it challenging to secure no-AI products and services. Innovation and security have not always been good bed fellows and AI is likely no different.

As AI develops if the tools are not secure then the output from these tools cannot be trusted; AI tools require access to data lakes held within behind organisations firewalls and the internet. If the AI is insecure you cannot guarantee that you can protect the data that it is using. You also cannot guarantee the quality of the output.

Here in lies a problem. When speaking with my audit colleague, you cannot attest to the quality of the output from AI if you cannot attest to the security of the AI. As you cannot guarantee that the AI model has not been changed. A U.S legal colleague told me recently that courts she deals with require lawyers to confirm if they have used AI to complete filings. If they have, they have to confirm that they have validated the AI results through traditional means.

@Thaddeus Dziekanowski AI and security go hand in hand.

Brian D. McCarthy Rachel V. Rose, JD, MBA

#cybersecurity #ai #cyberriskmanagement

View Original Post
https://www.linkedin.com/feed/update/urn%3Ali%3Ashare%3A7193504333223657472