Not too long ago, the notion of superhuman AI appeared far-fetched.

However, recent breakthroughs in the AI field have researchers now speculating that this might become a reality within the next decade, even though the exact timeline remains uncertain.

Given the rapid pace of progress, it has become increasingly imperative to delve into AI safety and regulation on national and international fronts.

Leading artificial intelligence companies, including Microsoft, Google, Amazon, Meta, OpenAI, Anthropic, and Inflection, have pledged to subject new AI systems to external testing before public release and to clearly label AI-generated content.

The White House recently announced these voluntary commitments as part of an effort to ensure AI systems and products are safer and more trustworthy, while Congress and the White House work towards comprehensive regulations for the rapidly growing industry.

President Joe Biden is scheduled to meet with top executives from these companies at the White House to discuss these commitments.

While some companies already perform internal “red-teaming” exercises to simulate potential issues with their technology, these commitments will be the first time all companies have agreed to allow outside experts to test their systems.

The external red-teaming will aid in paving the way for government oversight and regulation, potentially setting the groundwork for future external testing carried out by government regulators or licensors.

Watermarking AI-generated audio and visual content is another key aspect of these commitments, aiming to combat fraud and misinformation.

The companies have also promised to invest in cybersecurity and insider threat safeguards, particularly to protect the core knowledge base, known as AI model weights.

Additionally, they plan to establish mechanisms for third parties to report system vulnerabilities and prioritize research on the societal risks of AI.

Furthermore, they ‘intend to develop and deploy AI systems to address society’s most significant challenges’.

All of these commitments are voluntary, and White House officials acknowledge the lack of an enforcement mechanism to ensure companies adhere to them.

Critics, such as Common Sense Media, a child internet-safety organization, applaud the establishment of AI guardrails but caution that tech companies have a history of not fully committing to voluntary pledges.

They stress the importance of holding companies accountable for the impact of their products on the public.

The absence of early government regulation for social media companies has informed the White House’s approach to crafting AI regulations and executive actions.

Officials seek to make these commitments as robust as possible to avoid past mistakes and prepare for the rapid advancement of AI technology.

These commitments emerged from several months of discussions between the AI companies and the White House, which began with a meeting in May between AI executives and top government officials.

While these voluntary measures are a first step, the White House is also preparing executive actions and collaborating with lawmakers to develop comprehensive legislation to regulate AI.

In the meantime, the companies will begin immediate implementation of these commitments, with hopes that other companies will follow suit.

The White House anticipates that additional companies will recognize their responsibility to uphold safety, security, and trust standards and potentially join in on these commitments.

But as we already know, trusting Big Tech to do the right thing voluntarily is a risky proposal.

Real government-imposed regulation needs to happen.

Read the original article here:
CNN