Businesses are locked in a costly AI ‘arms race’

There is no doubt that we are in an AI arms race, says Jon Collins.

He has worked in the IT field for 35 years, holding various positions including software programmer, systems manager and chief technology officer.

He is now an industrial analyst for the research company Gigaom.

The current arms race has been spurred by the launch of ChatGPT in late 2022, Collins says.

Since then, many generative AI systems have emerged and millions of people use them every day to create art, text or videos.

For business leaders, the stakes are high. Generative AI systems are very powerful tools capable of digesting more data in minutes than a human could in several lifetimes.

Business leaders are suddenly realizing what AI could enable them and their competitors to accomplish, Collins said.

“Fear and greed are at the root of this situation,” he says. “And that creates an avalanche of momentum.”

With the right training, a custom AI system could enable a company to leapfrog competitors through a research breakthrough or by reducing costs by automating work currently done by humans.

In the pharmaceutical sector, companies are customizing AI to help them discover new compounds to treat diseases. But it is an expensive process.

“You need data scientists and model engineers,” says Collins.

These scientists and engineers must understand, at least to some extent, the pharmaceutical field in which the AI ​​will work.

And it does not stop there. “You need infrastructure engineers who can build your AI platforms,” he continues.

It is not easy to find such highly qualified workers.

There simply aren’t enough people who “understand how to build these systems, how to make them really successful, and how to solve some of the challenges that lie ahead,” says Andrew Rogoyski, director of innovation at the Surrey Institute for People-Centred HAVE. at the University of Surrey.

The salaries of those who can meet these challenges have reached “ridiculous” levels, he adds, because they are so important.

“We could produce hundreds of AI doctors, if we had the capacity, because people would give them jobs. »

Beyond the skills shortage, simply accessing the physical infrastructure needed for AI at scale can be a challenge.

The type of computer systems needed to run AI for cancer drug research would typically require between two and three thousand of the latest computer chips.

The cost of such hardware alone could easily reach more than $60 million (£48 million), even before the costs of other essentials such as data storage and networking.

Part of the problem for businesses is that this type of AI has emerged quite abruptly. Previous technologies, such as the emergence of the Internet, developed more slowly.

A large bank, pharmaceutical company, or manufacturer may have the resources to purchase the technology they need to take advantage of the latest advances in AI, but what about a small business?

The Italian start-up Restworld is a catering staff recruitment site, with a database of 100,000 workers.

Chief Technology Officer Edoardo Conte wanted to see if AI could benefit the company.

The company considered creating an AI-based chatbot to communicate with users of the service.

But Mr. Conte said that, for thousands of users, “the cost increases significantly.”

Instead, the study looked at a narrower problem: that candidates don’t always present their experience in the best way.

For example, a candidate might not list being a waiter among their skills. But the algorithms developed by Mr. Conte make it easier to discover additional information, including whether they have applied for and received a waiting role in the past.

“The AI ​​can infer that they are a server or that they might be interested in other server job openings,” he explains.

One of the obstacles to recruiting in the hospitality sector is getting candidates to the interview stage.

Mr Conte’s next challenge is therefore to use AI to automate and personalize the interview process for his candidates.

AI could even conduct a “conversation” with candidates and produce summaries to send to recruiters.

This could speed up the entire process, which currently can take days, during which time a server or chef could have found another job.

In the meantime, large companies will continue to invest money in AI projects, even if their goals are not always clear.

As Mr. Rogoyski says, AI adoption is in a “Darwinian and experimental phase” and it is difficult to predict the consequences.

“This is where it gets interesting. But I kind of think we have to go for it,” he says, before adding: “I’m not sure we have a choice. »