Learn more about Enabled Intelligence.
Let’s build American talent and jobs, and stop off-shoring mission-critical AI.
A blog from Enabled Intelligence.
From drones to hypersonic aircraft to advanced missile defense systems, AI technology is at the heart of solutions the US military is counting on to stay ahead of China and other adversaries. AI is not an end unto itself; what makes it a strategic capability is its potential to help maintain American weapon system superiority. China understands this. This is why the US must do everything it can to win the AI war.
What many people don’t fully understand is the importance of American human expertise when it comes to mission-critical AI. Just as it takes superbly trained warfighters to win a war, it takes human expertise to build, train, and deploy superb AI technology. And just as we should not rely on foreign gig workers to build, test, or operate American weapon systems, it makes no sense to rely on these same foreign workers when it comes to American AI technology.
However, while many proclaim to want the US to lead in AI, and at a time when we need to create American jobs for American tech workers, many AI firms in Silicon Valley rely on armies of low-cost foreign gig workers to build critical AI technology. While this approach may make sense in the short-term for some AI companies, it’s a very bad idea for national security.
First, there is the AI data quality problem. Every mission-critical AI solution relies on humans to annotate and fine-tune the training data it uses to generate results. You need high quality training data to go “in” the AI model in order to get “out” high quality results – for example, for AI to identify a Chinese missile in milliseconds, with 100% accuracy.
This means even small errors in AI data inputs can have catastrophic results for a weapon system that uses AI. Furthermore, poor quality AI training data increases AI compute costs and makes AI models difficult to deploy in the field, where the warfighter needs them.
And yet today, most US companies rely on AI training data prepared by foreign workers with no expertise in the data they are analyzing – while China, by contrast, has invested heavily in its AI workforce. American AI companies keep using low cost AI “sweat shops” and are then surprised when Chinese AI tech outperform them.
Relying on foreign workers also poses data security risks, and the creates opportunities for intellectual property theft. Furthermore, it undermines efforts to maintain a strong US military industrial base by diverting resources that could be used to train American workers – which means fewer jobs for US workers.
Fortunately, many US technology companies are now making substantial investments in hiring and training US workers. One example of this is Enabled Intelligence’s innovative geospatial data analyst training program, which trains US workers, including many veterans and persons with conditions such as autism to be highly productive AI data analysts. The success of these and other programs prove that using American workers leads to better AI.
It’s time for the entire US AI industry to embrace this reality.
/
