This Week in AI: US AI infrastructure order; Copilot AI monetization

Getting your Trinity Audio player ready...

Joe Biden signs executive order for development of US AI infrastructure

This week, U.S. President Joe Biden signed an executive order to expand and support the development of artificial intelligence (AI) infrastructure within the United States. The order focuses on building and scaling AI on U.S. soil to reduce the United States’ reliance on foreign infrastructure and to bolster national security, innovation, and the transition to clean energy.

“Advances at the frontier of AI will also have significant implications for United States economic competitiveness. These imperatives require building AI infrastructure in the United States on the time frame needed to ensure United States leadership over competitors,” says the executive order.

The order’s key measures include requiring the Departments of Defense and Energy to identify and lease federal land for AI data centers and clean energy facilities by 2027, align AI infrastructure with opportunities for clean energy generation, and boost domestic production of semiconductors and other components that are critical to the electric grid.

This executive order seems to be one of the few final efforts from the Biden administration to make a mark on the AI space before the Trump administration enters office. A few days before signing this executive order into effect, the Biden administration also implemented regulations that limit the export of U.S.-made computer chips to foreign countries.

The United States is currently the global leader in artificial intelligence, but China is a close second. This is why the U.S. implements policies like the two discussed above to give the United States the upper hand when it comes to AI innovation while stifling the competition and preventing them from receiving the crucial resources and software they need to surpass the United States as the leader in the artificial intelligence industry.

OpenAI unveils ‘Tasks’ on ChatGPT 

OpenAI recently launched “Tasks,” a new feature within ChatGPT designed to manage asynchronous tasks. In a demo shared by the company, Tasks was shown sending daily stock updates, exercise reminders, and summarizing news based on editable instructions provided by users.

Today we’re rolling out a beta version of tasks—a new way to ask ChatGPT to do things for you at a future time.

Whether it's one-time reminders or recurring actions, tell ChatGPT what you need and when, and it will automatically take care of it. pic.twitter.com/7lgvsPehHv

— OpenAI (@OpenAI) January 14, 2025

Tasks is currently in beta for ChatGPT Plus, Pro, and Teams subscribers and will eventually be available to everyone with a ChatGPT account.

However, after testing the feature, I find it underwhelming. When I tried to use Tasks, it fell short in execution on top of the significant limitations the feature already has. Although many people call this an AI agent, I don’t think it’s an AI agent, and I don’t think OpenAI wants us to think it’s an AI agent, so they deliberately avoided using that language when they released the product.

When I tried to set up a daily stock update through Tasks, it failed to deliver the updates automatically on its first day. When it finally worked, it sent notifications to my email without having the accompanying output in the chat interface. To be fair, the product has been out for less than 72 hours, but I was still surprised to see the system not able to initially handle the relatively simple task with which they demoed the product. However, since my initial test runs with Tasks, the issue seems to have been fixed.

While Tasks is definitely a signal that OpenAI is taking steps toward offering AI agents, Tasks feels more like a scheduling or notification tool at the moment. It lacks a degree of complexity that most end-users would want before they call the tool essential. Although Tasks clearly lays the groundwork for future innovation, the feature will need to become significantly more capable of genuinely impacting most users.

Microsoft’s 365 Copilot AI Monetization Strategy

This week, Microsoft (NASDAQ: MSFT) introduced Microsoft 365 Copilot Chat, which is essentially a chatbot that runs on GPT-4.0, has access to data stored in SharePoint to create more business-specific outputs, and includes a newly added pay-as-you-go option for its AI agent. I view this as Microsoft trying to open up a revenue stream on a product that is currently not profitable.

Microsoft’s strategy seems focused on onboarding businesses to adopt AI across their operations, coming for the free-to-use feature but becoming a customer via pay-as-you-go or by liking the free offering so much that they buy a monthly subscription for Microsoft’s AI offerings. I think they are trying to go this route because there is still a fundamental disconnect between how end users use AI products when left to their own devices and how the tech companies that have created these AI tools want their end users to use the AI offerings.

As Microsoft puts it in their official announcement, “Copilot Chat is a powerful new on-ramp for everyone in your organization to build the AI habit,” highlighting that, for many people, using AI in the workplace isn’t a habit yet.

Most end-users still interact with AI in a very primitive way, using it like a search engine. This limited capability—usually presented as a free-to-use chatbot—does the job for many people, which means many people don’t have an incentive to upgrade to a premium model. As a result, AI companies are struggling to profit from services they’ve invested heavily in because they are shelling out billions of dollars per year to train and run their AI models while lacking the number of paying customers needed to cover those costs. 

Microsoft’s strategy seems to align with a previous prediction that I made when Microsoft first announced its AI agents: that there needs to be a re-education campaign to help users understand how these tools can be used in more sophisticated ways–which could eventually lead to tech giants like Microsoft creating revenue streams for more of their AI offerings.

In order for artificial intelligence (AI) to work right within the law and thrive in the face of growing challenges, it needs to integrate an enterprise blockchain system that ensures data input quality and ownership—allowing it to keep data safe while also guaranteeing the immutability of data. Check out CoinGeek’s coverage on this emerging tech to learn more why Enterprise blockchain will be the backbone of AI. 

Watch: Turning AI into ROI