Recommended: Listen to this story (00:49 - 08:10):

The Story: Last week, president Biden signed an executive order called “Safe, Secure, and Trustworthy Artificial Intelligence.” It’s the first ever federal guardrails given for AI by the United States government. And it’s a sweeping order—according to Bruce Reed, the White House deputy chief of staff, this is the “strongest set of actions any government in the world has ever taken on AI safety, security and trust.”

But that doesn’t mean those actions will necessarily be effective. At least according to Richard Dulude, co-founder and general partner at Underscore VC.

Not long after Biden’s executive order was announced, Dulude posted on LinkedIn, outlining the things he believes the order got wrong.

Dulude writes, “The recent order will be looked at as a marquee motion in the history of technology regulation. After a read and discussion with some of the smartest folks out there in the AI community today -- here is what I surmise they got wrong, and what we could still get right in the decades ahead 🇺🇸:”

Dulude follows that with these three observations:

1. We need to regulate *applications* of AI, not foundational models.

2. We need to focus on *industry-specific evaluations* of AI, not general evaluations.

3. We need to ensure that *open-source AI* has a fair opportunity to succeed, by not protecting closed alternatives.

Dulude ends his post with this: “Discrimination, fairness, the muddling of truth, and the right to privacy are all on the frontlines of the debate, and what is certain, is that bright enterprising founders are going to build products that solve these problems.”

Expert Take: We spoke with Richard, co-founder and general partner at Underscore VC, for some elaboration on his LinkedIn post.

Dulude believes the executive order misses the mark in neglecting to establish rules in an “industry-specific way by an industry-specific body.”

He goes on, “The [executive order] just doesn’t seem the right approach. If you really want to get to what is an understanding of truth and how these models should be managed, let the biggest and broadest audience input into that assessment of what is true and what is fair.”

*Stay informed about the three biggest stories in venture capital and tech news every weekday morning. 2-min reads only.