Also, biological weapons‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌  ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌ ͏‌  ͏‌ ͏‌ ͏‌

John,

One of the best parts about being a member of Congress is that - if all you want to do is learn - you can get almost anyone in the country on the phone.

So, for the last few months, I’ve used my position on the Science, Space, and Technology Committee to learn as much as I can about artificial intelligence.

I’ve given my staff the names of people I wanted to speak with in the A.I. community - industry leaders, academics, think tanks, national security folks - and I’ve had the privilege of dozens of conversations.

Here’s some of what I’ve learned:

A.I. is an enormous tidal wave and it’s going to make landfall years before we thought it would.

The entire tech sector is sprinting toward A.I. because they think the winner gets to stay in business and the losers don’t. And while the business leaders I’ve spoken to like to talk about all the innovation that’s happening, I can always hear an anxiety underlying all of it, namely, that they know this is now the Great Race and they’ve got no choice but to start running.

As a result, we’re being pushed into a full-scale technological revolution.

You’re about to see A.I. everywhere, including politics. There will be lots of high-quality deepfakes in this election.

There are safeguards being developed to guard against misinformation - I’ve talked to some groups that are working on them - but they’re in the early stages and I don’t think they’ll be ready in time for this election.

Which means you have to be really careful with video clips on social media - like mine. There’s not a great solution right now other than your skepticism.

In contrast, I’ve also heard concerns that the regulation of A.I. could go too far and block a lot of innovation.

But that’s not going to happen, and I’ll tell you why:

The political will in Congress to do some big regulation of A.I. just isn’t remotely there. The vast majority of Congress has never used ChatGPT and most of them are only vaguely aware of what it does. We’re starting to see a hazy consensus that “A.I. is an issue,” but that still puts us miles away from, “And here is the solution - and here are the votes for that solution.”

Just to be totally candid with you, the much greater chance is that Congress ends up doing basically nothing to establish real safeguards - just as we failed to do with the internet and social media and cryptocurrency.

Acknowledging the historical tendency toward congressional inaction is really important in forming a legislative strategy. It’s helpful to start from a place of, “Any bill in this space will likely fail, so how do we defy the odds?”

To me, this means focusing on the most important thing we absolutely have to get right, which I believe is preventing the intentional misuse of A.I. - like creating new biological or cryptological weapons.

Some of the conversations I’ve had with respect to those matters have been quite chilling. The basic problem is not that ChatGPT could be used to devise new weapons - the company that created it has put lots of safeguards in place, as have the handful of other domestic companies who have built similarly powerful A.I. interfaces.

Rather, the concern is that other bad actors could create their own A.I. without those safeguards and could use it to guide the engineering of new weapons.

Basically, if you buy 20,000 GPUs (a specific kind of microchip), link them together, and give them about nine months to train on an extremely large data-set, you now have an incredibly powerful A.I. that you can use to wreak havoc.

Today, doing all that would set you back about $200 million, so we’re basically talking about adversarial nations at this point.

I’ve got a couple of bills to address that concern. Getting them passed is going to be a challenge, but I can share what the first steps have been.

After I got them drafted, the next step was to find a member of the other party to join me. (As you can imagine, it’s very important to have a co-sponsor from the majority party if you want to get something passed.)

I started by going down the list of members of Congress who are in the majority party and scanning for someone I thought would be interested in this. I picked someone I hadn’t met, but whose background and current committee assignments looked like a good fit.

The next day, when we were all summoned to the House floor for a vote, I literally walked across the aisle, found the member, introduced myself, and told him about one of my ideas for a bill. He told me he was interested, so we’ve been working on it and I think it’s about ready for us to file it. More on that soon.

And that’s what bipartisanship looks like in practice. Hopefully, I’ll have good news for you in a couple weeks.

Congressional Baseball Game

On a completely different note, last week I told you I was about to play in the congressional baseball game and would report back.

It was my first time in a major league baseball stadium in my life. (I’m from North Carolina and we don’t have a major league team, so we’d root for the Braves but I never made it to a game.)

Walking into a major league stadium for the first time as someone who was about to play was pretty great. Marisa was there with me, as was most of my staff.

20,000 people attended and we raised about $2m for charity.

I was used as a pinch runner. I was brought out three times and scored twice. Stole third base, as you see here:

Jeff sliding to a base on the baseball field.Photo Credit: Bill Clark, CQ Roll Call

Aaaand we lost, 6-16. But last year we lost 0-10, so at least things are trending in the right direction.

Next week, I’ll have the story about what it was like to spend all day in a room with the Armed Services Committee passing the annual defense bill. Finished yesterday at 1:00 a.m.

Best,

Rep. Jeff Jackson (NC)