Silicon Valley presented AI as a noble research tool. Now it's all about cold, hard cash.

• Tech firms once had the luxury of pursuing expensive AI research.
• But the rise of ChatGPT has pushed them to prioritize the commercialization of the technology.

There was a time when companies working on artificial intelligence seemingly had the luxury of experimenting with any research endeavor of their choosing – without any worries about the bill.

You only need to look as far as the output of AI lab DeepMind in the near-decade since its acquisition by Google to see that's been the case: AI systems figuring out protein folding, fusion energy, or games like Go had all kinds of time and resources poured into them.

For researchers like those at DeepMind and elsewhere, being given the time to pursue potentially groundbreaking research without a second thought to the cost of it all has been freeing. In 2020, Google wrote off $1.3 billion of DeepMind's debt in a sign of its commitment.

But in the ChatGPT era, it's becoming pretty clear that priorities are being completely flipped. For tech leaders, now it's all about making cold, hard cash.

Since the launch of OpenAI's chatbot, a race to commercialize AI has gathered pace across the tech sector as the consumer interface of ChatGPT has galvanized users on the advantages of AI while prompting competitors to come up with a viable alternative or risk losing business.

The latest instance of companies prioritizing returns comes from Meta. The Financial Times reported that the company axed a team involved in protein folding research as part of its massive restructuring program, which Mark Zuckerberg has called the "Year of Efficiency."

What's notable is that this division has been axed despite AI now being Meta's top investment; though it's a niche field, the use of AI in solving protein folding is an important area of research that could help scientists understand all kinds of diseases like Alzheimer's and Parkinson's.

What Meta seems to be prioritizing instead is stuff that could potentially make them money. During Meta's recent earnings call, Zuckerberg told analysts, for instance, that AI-recommended content had helped boost time spent on Facebook by 7%.

That, in turn, boosts money earned from ads, which constitutes the vast majority of the $32 billion in revenue Meta made in the three months to the end of June.

The pressure to make money has been seen elsewhere, as Google formed a new combined unit with DeepMind in April that delivers AI research papers that "dramatically improve the lives of billions of people" and products too.

It is worth noting that last month, Meta did release Llama 2, its new large language model, as a free-to-use, open-source tool for businesses with fewer than 700 million monthly active users.

The company's vice-president of AI, Joelle Pineau, also said Meta "remains committed" to its Fundamental AI Research (FAIR) Team, which conducts "exploratory research based on open science." Meta did not immediately respond to Insider's request for comment made outside of normal working hours.

But Meta is making moves to commercialize AI in as many ways as it can. As the Financial Times reported earlier this month, the company is preparing to roll out consumer-facing chatbots much like ChatGPT as soon as next month.

The race to make money from AI then is heating up. Everything else is likely to cool down.