### Future of AI Regulation under Trump Administration

The incoming administration may prioritize innovation while potentially rolling back civil rights protections in AI regulation.

### The Future of AI Regulation: Anticipating Changes Under a New Administration
* 🤔 The focus on American AI regulation raises questions about the **potential repeal of Biden's executive orders** by the incoming Trump administration.
* 🔄 While Trump may attempt to undo some of Biden's initiatives, much of the groundwork has already been laid by various agencies, making **complete reversal unlikely**.
* ⚖ The incoming administration is expected to target specific areas, particularly around **equity and civil rights**, which are significant to the Republican base, potentially leading to **concerns about biases in AI systems**.
* 📉 Companies like Amazon and Apple have faced scrutiny for biased AI tools, and while they may prefer **self-regulation**, the inconsistency in approaches could lead to **inefficiencies**.
* 🚀 The Trump administration's first executive order on AI R&D is seen positively, and there is hope that a similar focus on **innovation** will continue, especially with key figures from that era involved in **planning**.
* 🌍 The administration is likely to prioritize **accelerating U.S. innovation** while strategically managing competition with China, including **extending export controls** established by Biden.
* 📊 Overall, the regulatory landscape may shift towards a **balance between fostering innovation and addressing safety concerns**, requiring careful consideration of the implications for AI development.

What's interesting about you is you
particularly focus in on Americans, fact focus your senior vice president of
government affairs for the Americans for Responsible Innovation, responsible
innovation, particularly around AI, and many would say actually the steps taken
by the Biden administration, whether it

was executive orders put in place, some
sort of guardrails, will they be taken down, Doug, do you think?
That's a great question, Caroline. It seems likely that incoming President
Trump will likely repeal the Biden executive order.
It's more of a question about whether he

can undo some of the things that
President Biden has done. A lot of the work that has been sent to
the agencies has already been completed. They've taken a lot of tasks and done a
lot of research pursuant to the executive order, and that doesn't go
away overnight.

Even at the university itself is remote.
Okay. Why don't they go away overnight?
They've already been in action to a further extent, Doug.
That's right. And President Trump, if he wants to, he
can undo some of the work that's been

done.
He can scrap some of the secret signifiers, some of the people.
But a lot of the work itself has already been done.
And there's the reasons that we believe that the incoming administration may
want to keep some of these processes in

place.
You know, by all accounts, the agencies have done professional work.
A lot of it is non-ideological. It's just understanding what AI systems
are capable of and how the agency's mandate fits into that.
So it seems unlikely that he's going to

try to pull everything out, root and
branch from the beginning. It does seem like he will pick a few
items that he really cares about and try to undo that work.
And let's just talk about what therefore could be undone and whether that's
actually the right balance in terms of

pro-innovation rather than stifling
safety. Sure.
It seems more likely than not that the incoming administration is going to undo
some of the work that the Biden administration had done around equity
and civil rights. I think those topics are very important
to the Republican base, and I think that

any time the word equity or civil rights
appears in the executive order and the agencies work, the new administration is
probably going to try to unwind that. What does that mean in practice?
Ultimately, that makes us all worry about the biases that get put into the
systems.

Or do you think that there will be some
manner of self-regulation within the companies that are building these all
powerful large language models now? It is challenging, and I think a lot of
the companies, they welcomed these actions because, you know, we all
remember some of the headlines a few

years ago.
I think immediately of, you know, Amazon had an internal recruiting tool which is
found to favor men over women because more of the existing employees at Amazon
were men. And to their credit, they recognized
this and didn't release it at all.

There's other stories about I believe
Apple had a credit card and at some point it was favoring men over women
because of the way the data was trained. I mean, these are terrible headlines for
any publicly traded company, and they

don't want that.
I do think even the federal government itself is not going to do this research
and test the systems for these biases. The big companies, particularly the
major publicly traded companies, are going to want to do this on their own.
And I think there's a lot of, you know, 

there's a lot of duplication and
inefficiencies, and every company has to do it themselves rather than have better
government do it centrally. But I think that's probably what we'll
see. Okay.
Well, what's interesting is if you go

back to the first Trump administration
was actually an executive order there that focused R&D on AI in many ways.
So what do you think from a positive perspective could be done in a new
administration that does perhaps open the world a little bit more to
innovation, even if it puts certain

worries about bias, civil rights, not
within the federal jurisdiction? That's right, Carolina.
And, you know, at Emory, we're fans of the first executive order that President
Trump did in the waning days of his

administration.
The man who led that work, someone named Michael Kratsios, is, by all accounts
playing a major role in the Trump team's planning for tech policy in the next
four years. And we're excited to see where that
leads. 

I think that the first executive order
went a long way towards, you know, encouraging agencies to understand where
AI fit into their work and take proactive steps to try to get ahead of it.
And that's something that the Biden administration continued.
We think that the Trump administration

is likely to continue this, particularly
if someone like Mr. Kratsios is in charge of the policy.
I think beyond this, you know, we've talked a little bit about China and what
that means. I think the Trump administration is
likely to go full speed ahead on all

things to try to accelerate US
innovation and try to hold back China. But there's a lot of policies in the air
space that is not about simply cutting regulations.
You need to be smart. A lot of the export controls that the
Biden administration has done, the Trump

team could extend.
They could try to prevent open weight models from making it to China.
And these are all serious regulatory efforts that go well beyond simply
cutting regulations and letting industry do whatever it wants.

Trump's Future for AI

Welcome to the Bloomberg Technology YouTube channel. Subscribe for original series and mini-docs on game-changing technologies, profiles on innovators, live conversations with experts, and breaking technology news & analysis from Bloomberg TV.

Bloomberg Technology

Potential Reversal of Biden Administration's AI Policies under Trump

Likely repeal of Biden's executive order on AI.

Difficulty in completely undoing Biden's actions due to completed agency work and research.

Trump could potentially undo some completed work, but not entirely.

Reasons for the incoming administration to maintain some processes due to the professional and non-ideological nature of the agency's work.

Focus on specific items for potential reversal, considering the balance between pro-innovation and safety.

Concerns Regarding Equity and Civil Rights in AI

Likely reversal of Biden administration's work on equity and civil rights due to Republican priorities.

Worry about biases in AI systems due to potential unwinding of equity and civil rights initiatives.

Discussion of potential self-regulation within companies developing large language models.

Examples of past biases in AI systems from companies like Amazon and Apple.

Expectation of self-regulation by large companies to avoid negative publicity.

Potential Positive Impacts of a New Administration on AI Innovation

Discussion of the first Trump administration's executive order focusing on AI R&D.

Positive view of the first Trump administration's executive order and the potential role of Michael Kratsios.

Expectation that the Trump administration will continue encouraging agencies to understand AI's role.

Focus on accelerating US innovation and countering China.

Potential extension of export controls and prevention of open-weight models reaching China.