Page 55 - Cyber Defense eMagazine February 2024

The implementation of regulations coming out of the order are poised to make changes in the future, be
it the pace of business cycles, types of solutions and offerings, business workflows or even expansion of
new revenue streams.

AI-volution and the Future

There has been a steady shift in certain processes, particularly as it concerns safety and flagging off
content. Traditionally, issues regarding safety and flagging off disparities like biases would be done in
retrospect, after AI applications were developed. While this approach does help identify the discrepancies
and implement learnings in the future, it doesn’t solve the current problem.

Responsible AI by design, as an approach, weaves in transparency, fairness and accountability right
form the design phase. It focuses on how we can integrate responsible AI principles at the time of
development of AI apps or foundational models, not after the process. It enables training the algorithm in
such a way that they adhere to the regulations from the start.

Anthropic is a good example of a company that’s scaling AI safely and responsibly with intentionality.
They have also instituted a responsible scaling policy (RSP) aimed at mitigating catastrophic risks, in
tandem with other measures to mitigate bias, toxicity and misinformation, protecting customer privacy,
building robust and reliable systems and using constitutional AI. In fact, their tool Claude 2.1 has made
significant gains in honesty, with a 2x decrease in false statements compared to the Claude 2.0,
improving the overall accuracy of the tool for its users.

Both regulation and intentionality are crucial to ensure that AI evolves in the right direction. The EO uses
a combination approach of guidelines and regulation, encouraging companies to develop responsible AI
applications and models.

The future of AI is bright, but also uncertain. President Biden's executive order lays a framework to steer
this powerful tool in a direction that maximizes business and societal benefit while minimizing harm.
Regulation and intentional development focused on responsibility are crucial to ensure AI evolves safely
and for the greater good. There will always be risks, but if we act thoughtfully the benefits will far outweigh
them.

Cyber Defense eMagazine – February 2024 Edition 55
Copyright © 2024, Cyber Defense Magazine. All rights reserved worldwide.