Hello and welcome to Eye on AI.

Beware, cybersecurity professionals, AI is coming for your jobs—or so the theory goes. 

The rationale is something like this: AI tools will eventually handle a lot of the grunt work involved in protecting computer systems, like automatically scanning files for suspicious behavior and writing summaries of its findings. Therefore, there will be less work for humans and possibly layoffs.

But Steve Schmidt, Amazon’s chief security officer, disputed the jobs-killing argument and said there’s more than enough cybersecurity work to go around. 

“I would be astonished if AI drives layoffs in the security industry,” he told Fortune.

That’s because of the big gap in available talent and the number of open jobs in the security industry right now, said Schmidt, one of Amazon’s famous “s-team” executives—or senior leaders—who reports directly to CEO Andy Jassy. 

The global cybersecurity workforce headcount of 5.5 million is at a record high, according to ISC2, a nonprofit that trains security professionals. But even that’s insufficient to meet industry demands. Companies worldwide need 4 million additional cyber professionals to secure their systems, a number that has grown annually since 2021 when ISC2 estimated the shortage at 2.7 million. 

It’s not just Amazon that is unlikely to cut cybersecurity jobs due to AI. Other chief security officers Schmidt has spoken with told him they won’t cut their staff either because of the technology, according to Schmidt. They will just get more done with AI’s help, he said. 

The optimism comes as layoffs have plagued the security industry over the last year, hitting both cyber companies and cyber teams within businesses. Five U.S. security giants—Secureworks, Rapid7, Proofpoint, Splunk, and Okta—have cut 1,880 workers combined since August. Most cited cost reduction as the reason for the layoffs. None of them pointed the finger at AI. 

In the larger tech world, Meta’s layoffs of 21,000 employees since late 2022—part of an effort to make the company more efficient—included some workers focused on website security, though the exact number is unclear. Microsoft also cut “hundreds” of U.S.-based security employees last year due to the economic downturn, Business Insider reported. 

Whether or not AI played a role in these layoffs, the fear about the technology taking over cybersecurity jobs has some basis. Human labor is expensive, and the salaries of security engineers have increased in recent years due to the shortage of talent, said Unal Tatar, assistant professor of cybersecurity at the State University of New York at Albany who has worked in the field for two decades. 

“The major budget line is people,” he told Fortune. “AI never gets sick. It doesn’t take time off, and it can run 24/7.” 

That’s not a problem at Amazon, though, according to Schmidt. Speaking about AI tools, he said, “I don’t see them replacing the really skilled security engineers that I have.” 

Cyber security is playing chess on a unicycle

Schmidt described his job as solving a puzzle, playing chess, and practicing psychology at the same time. AI tools can lighten the load for security engineers, but as hackers also get their hands on the technology, it can mean adding a new level of difficulty—unicycling or knitting, for example—to Schmidt’s juggling.

Just as engineers can use AI to do their jobs better, adversaries can use AI to quickly and more efficiently hack into computer systems. Large language models can help write more persuasive phishing emails, for example. Hackers can also use AI to invent new ways to take advantage of people, like cloning the voice of a loved one and asking for money over the phone. 

“Our industry is one where there is no lack of opportunity to improve,” Schmidt said. “It is always a race.”

The biggest beneficiaries of AI will be small- and medium-sized businesses that can’t afford big teams to create and maintain sophisticated defenses. For example, companies creating their own software can use AI products to review their code and suggest ways to make it more secure. 

“Where we will see some phenomenal jumps in the not-so-distant future is empowering people who don’t have the security teams that I do, who don’t have the level of expertise that my team has,” he said. 

Amazon has a generative AI assistant that helps clients write code, among its many AI products. It also sells Bedrock, a service that helps companies build custom generative AI based on existing large language models. And Amazon, of course, benefits when companies buy its services. 

Tatar agreed that AI won’t cause any significant layoffs in the security industry, at least in the short term. The technology isn’t ready yet, he said, but even when it is, it will take time for customers and investors to support its adoption. 

Where AI tools could have an impact on the workforce, however, is with entry-level jobs, Tatar said. Even now, early-career roles are few and far between because many open jobs require years of experience. If AI begins to take on the job responsibilities of early career security analysts, he said, it could become harder for new college graduates to find work.

Please let me know what you think about the future of cybersecurity careers.

With that, here’s more AI news.

Rachyl Jones
[email protected]


Trouble in the ranks of Google DeepMind. Alphabet last year merged its two AI research units, Google Brain and DeepMind, under DeepMind founder Demis Hassabis. But the two groups already had tensions between them, and Hassabis is unhappy with the arrangement, according to new reporting from The Information. After being allowed a loose leash to pursue AI research at DeepMind, Hassabis now faces pressure to produce products that Alphabet can sell. Hassabis is frustrated with the change, sources who spoke to The Information said. In the article, a DeepMind spokesperson denied any tensions between the AI research teams and denied any frustration from Hassabis. 

The marketplace for AI training data is booming. Big Tech companies are racing to buy vast troves of photos and videos from secret suppliers to train their AI models, Reuters reported. The $2.5 billion market of buying training data could grow to $30 billion in a decade, Business Research Insights said. Photo services Shutterstock and its competitor Freepik have licensed their images to some AI companies, and image hosting site Photobucket is in discussions to do so, according to Reuters. News organizations including the Associated Press and Axel Springer have also made licensing agreements with OpenAI while the New York Times has sued OpenAI and Microsoft over the matter. 

When companies can’t buy training data, they may skirt rules to take it. Companies including OpenAI, Google, and Meta have ignored internal policies and considered breaking the law to get their hands on the massive amount of data necessary to train their AI models, a New York Times investigation found. According to a Google internal message, a change to the company’s terms of service last year was partly motivated by gaining access to user data in Google Docs and restaurant reviews on Google Maps to develop its AI products. OpenAI and Google have both transcribed YouTube videos to train their AI models, the Times found, which could violate copyright laws. 

Meta discussed buying publishing house Simon & Schuster. The social media company last year considered buying the book publisher to access its library and train Meta’s AI models on the writings, the Guardian reported, citing a New York Times investigation. Meta had reportedly used almost every English book, poem, and essay on the internet to train its models and needed more content. The publisher sold to private equity firm KKR in August.

AI is the diamond of the century, Dimon says. Jamie Dimon, the JPMorgan Chase CEO, told investors in an annual shareholder letter on Monday that AI will have as big of an impact on society as the printing press, electricity, and the internet. The bank employs more than 2,000 AI and machine learning experts and sees more than 400 use cases for the technology, including in marketing and fraud detection, he said. Read more about the letter with CNBC


Will AI reduce credit quality for tech companies? AI-related revenues will reach nearly $650 billion in the next four years, S&P Global Ratings said in a report published Monday. The semiconductor industry is the breakout winner for the increased spending across the AI sector, and hardware companies are beginning to benefit as well. But it will take some time for the software and services sectors to begin generating meaningful revenue, S&P Global Ratings analyst Andrew Chang said in the report. Many credit issuers will benefit from the investment cycle, but some will be “disrupted,” the agency said. 


AlphaSense, a Goldman Sachs-backed AI research startup valued at $2.5B, gears up for IPO as it crosses $200M in annual recurring revenue —by Jessica Mathews 

Open source AI is booming, but OpenAI’s GPT-4 is still the big winner with corporate customers—for now —by Sharon Goldman

Elon Musk’s friends are reportedly looking to help raise $3 billion for his AI startup as it plays catch-up with rivals OpenAI and Anthropic —by Jason Ma

‘Social order could collapse, resulting in wars’: 2 of Japan’s top firms fear unchecked AI, warning humans are ‘easily fooled’ —by Chris Morris

Top portfolio manager at a $194 billion asset manager breaks down her top 3 AI stocks for the next decade—and one honorable mention flying under the radar —by Will Daniel

Vinod Khosla is betting on a former Tesla autopilot engineer who quit to build small AI models that can reason —by Sharon Goldman  


April 15-16: Fortune Brainstorm AI London (Register here)

May 7-11: International Conference on Learning Representations (ICLR) in Vienna

May 21-23: Microsoft Build in Seattle

June 5: FedScoop’s FedTalks 2024 in Washington, D.C.

June 25-27: 2024 IEEE Conference on Artificial Intelligence in Singapore

July 15-17: Fortune Brainstorm Tech in Park City, Utah (Register here)

Aug. 12-14: Ai4 2024 in Las Vegas


$11.6 billion 

That’s how much the U.S. government will give Taiwan Semiconductor Manufacturing Company, the world’s most valuable chipmaker, in grants and loans to build semiconductor factories in Arizona, according to Bloomberg. The investment is part of President Joe Biden’s efforts to ramp up chip production in the U.S.—chips are a necessary component to AI growth—as China makes headway on its own semiconductor industry. The Taiwanese company will receive $6.6 billion in grants and up to $5 billion in loans.