JAIC director Shanahan on how to build trust with Congress

I have struggled with this at the JAIC going up to the Hill and saying, ‘You know, you’re asking for a level of detail in fiscal year 2023 on what I’m spending on in technology. I’m not certain what six months is looking like, I cannot promise you’ — and in fact, I’m not really being tongue in cheek or glib about this — I say, ‘If I tell you exactly to the dollar amount what I’m doing in fiscal year 2023, you ought to fire me because I’m lying to you. I don’t know how fast technology is going to adapt between now and then.’

 

That’s incumbent upon us to tell the right story. It’s the same thing I know Will Roper is trying to do on JADC2 and other are trying to do. Army Futures Command has that same dilemma, ‘How do I translate PowerPoint-deep ideas into trust me.‘ And that’s hard. Sometimes the Department doesn’t have the right track record on the trust side. What I always say is ‘Hold us accountable. If I can’t tell you what that dollar amount is, but you can hold me accountable for delivery of product,’ — but beyond that is a ground game of going up there and really engaging with staffers and members over and over again to get them comfortable. We haven’t been great at that in the Department of Defense.

 

We have a former staffer now and she says we’re nuts to avoid going up there and treat them like the adversary. These can be our friends in court. I’ve never had a more bipartisan issue in my time in uniform. Everybody wants the DoD and AI to succeed. And I think it’s the same thing with JADC2, but it’s show me something. I think we’re going to have to show it incrementally, build little victories, while still talking about the grand vision, but then bring in people from the commercial world who have credibility in that world and say, ‘No, this is the right approach, and let me tell you what we’re going to do over the course of the next year, and I’ll tell you what, we’ll come back to you with IPRs [integrated program reviews] every six months, every four months, or every two months,’ and build that relationship with Congress that sometimes hasn’t been as strong as it needs to be.

 

They do need to be convinced, and it’s incumbent upon the Department to get them there. I can’t blame them for asking really hard questions when they keep hearing this term AI or JADC2 and really want to see what its about.

That was the outgoing Lt Gen Jack Shanahan, director of the Joint Artificial Intelligence Center (JAIC), speaking with David Deptula on the Aerospace Nation series. A lot of good stuff throughout the episode, listen to the whole thing!

I think Shanahan is right on with most of that statement, but one of the hard parts of that is holding defense leaders accountable. For example, Shanahan is already on his way out the door after 18 months at the JAIC. So clearly he cannot be held accountable for the choices he made in the past.

A related issue is that he said he cannot predict exactly what he needs to be doing two years into the future to justify budgets. That’s exactly right. The JAIC is interesting because it actually has something of an organizational budget I’ve been advocating for, allowing it to move money to new things in the year of execution (i.e., be agile). The JAIC is the program, not a particular project with specified and approved requirements. But Shanahan is also saying he needs to communicate much more with Congress, which seems counter to his point that budgets can’t be specified.

What I think he’s saying is this: Give the Department more flexibility in defining its budgets, which due to long lead times cannot be specified exactly two years before. Attention should shift from line item justification in the budget to justification of what will be done with their funds in the year of execution. Continually communicate with Congress that, with this $132 million in RDT&E for the JAIC, I’m planning to send it here, here, and here in the upcoming quarter, and these are the reasons why and what you’ll get for it. Also, these are the things we’ve been working on and we can now demonstrate the value in an iterative, real-world way, rather than promises from two years ahead of time.

So again, fund an organization to a specified level with a general justification of the types of missions and activities. Then, spend far more time in the year of execution detailing to Congress where the funds are going and to what projects. And finally, spend even more time showing Congress what has been accomplished with iterative MVPs and real-world tests from the projects authorized from the past year. That’s how you build trust. Failure to communicate and failure to show delivery of capability is then cause for slashing budgets or removing managers.