James Bond's job is safe as GCHQ scientist says AI can only spy on 'very young'

The findings suggest that James Bond and his colleagues have nothing to fear from the appearance of machines. Photo: Nicola Dove/PA

“At an early stage of writing the report, the role of large language models can be determined, akin to treating a large language model as an extremely junior analyst,” they said.

Large language models work by consuming and parsing huge amounts of data, such as text from the Internet, and using it to write an answer to a question, determining which word should be next. However, they do not understand the meaning of what they write. As a result, they often give incorrect answers and seem to come up with false «facts».

Adam S. and Mr. Carter, in their article titled «Large Language Models and Intelligence Analysis» say: «Performance Assistant»; auto-complete sentences, check emails, and automate certain repetitive tasks.”

«The most promising use cases are still on the horizon, and future efforts should focus on developing models that understand the context of the information being processed, rather than just predicting what the next word will be.»

< p>Chatbots with artificial intelligence, such as ChatGPT and Google Bard, are described as «a team member whose work, when properly supervised, has value, but whose products will not be released … without significant review and verification.»

Fears that AI will replace humans in the workplace are fueled by research showing which sectors are most likely to be applied to AI. One study by McKinsey predicted that 12 million jobs could be lost in the US alone over the next seven years due to its spread to the corporate world.

AI tools are increasingly being introduced into the workplace following the public launch of ChatGPT late last year.

Is artificial intelligence developing too fast?

Google is using AI chatbot technology to process new hires, and Microsoft is using the technology to provide technical support to its employees.

While the work of senior intelligence officers is not yet at risk, security officers expressed concern about the potential impact of AI on democracy through the mass production of disinformation.

Adam S. and Mr. Carter stated: “The ability of a statesman or organized crime group to launch disinformation campaigns has been vastly improved with…large language models. But what's more worrying is that [they] have now allowed less sophisticated actors and opportunists to potentially do significant damage, thereby lowering the entry barrier for nefarious actors.»

Senior intelligence officials have also noted the threat poses AI as a tool for cyberattacks and espionage. Shortly after the public release of ChatGPT, GCHQ warned that chatbots were a new security threat.

Non-tech companies have heeded these warnings. The city's law firm, Mishcon de Reya, has banned its lawyers from using ChatGPT due to concerns that sensitive material could be leaked online. Investment banks, including JP Morgan, have imposed similar restrictions for comparable reasons.