Should Spies Be Worried About AI Taking Their Jobs? Head of UK’s MI6 Says No – Here’s Why

The Director of MI6 has assured the agency’s spies that the rise of Artificial Intelligence(AI) does not threaten their roles. Instead, the agency is already employing AI to assist the agency and its personnel in its operations.

MI6 Director Assures Spies of Their Vital Role

The loss of jobs to AI has been one of the main causes of AI anxiety since the technology went mainstream. According to the World Economic Forum’s Future of Jobs 2023 report, however, this is not a naive concern. The report predicts a 44% disruption in roles in various sectors as well as a massive loss of jobs in others such as administrative roles.

According to a speech delivered by Richard Moore, the director of UK Intelligence, MI6, spies, and espionage are not part of those who will lose their jobs to the technology. “AI is going to make information infinitely more accessible and some have asked whether it will put intelligence services like mine out of business. In fact, the opposite is likely to be true,” he said.

Moore added that the unique characteristics of human agents in the right places will become even more significant. He termed his spies to be more than just collectors of people’s data. They can also be tasked and directed and can identify new questions that the agency had not thought to ask.

More importantly, they can sometimes influence decisions inside a government or terrorist group.

He, however, acknowledged the potential that AI would advance to a point at which it could “overtake some aspects of human cognition.” He additionally noted that it might occur more quickly than was anticipated.

Still, he maintained his belief that AI would not be able to take the place of the “extraordinary bond” that exists in interpersonal interactions, which is the foundation of gathering intelligence.

”However swift and all-encompassing the advance of AI, some relationships are going to stay uniquely, stubbornly human, and those relationships are at the heart of my service because my agency is dedicated to preserving human agency.”

However, MI6 is not ignorant of the potential of the technology to amplify and enhance their work. While it will not replace human spies just yet, Moore stated that it is being used to augment their operations and decision-making.

Sir Richard Moore head of MI6 declares AI will help not hinder agents.
China, on the other hand, leader in AI, says AI is potentially dangerous.

— Roberta Sutton (@Roberta9996) July 19, 2023

More precisely, Moore revealed that MI6 was using AI to disrupt the flow of weapons to Russia for use against Ukraine. Using AI and loads of data, the agency is working to stop the attack going on in Ukraine by interrupting and eventually halting the supply of ammunition to Russia.

Embracing AI in Intelligence Service

AI is useful in espionage activities because, like spying, AI’s main function is also to identify patterns. With access to reliable training data, it can learn and generate patterns at a much faster rate than human beings.

Additionally, AI can support and accelerate the process of analyzing information, identify trends, and eventually assist in decision-making. According to the Government Communications Headquarters (GCHQ), AI assists human investigators by capturing and intercepting images, messages, and chains of contacts to help identify perpetrators.

However, the adoption of AI in spying has been faced with difficulties. This is mainly due to the importance of Intelligence agencies’ reputation and how an error or failure caused by AI could ruin it due to its shortcomings.

In most cases, spying entails finding outliers and extremists among many average and normal activities or persons. Unfortunately, AI performs better in average cases which fits the description of most of its applications except intelligence operations.

This makes it less efficient when applied to such operations and increases the potential for error. Considering how important spy operations tend to be, the error consequently affects the decisions made by the agency as well as the government.

Secondly, AI is less trustworthy in complex and dynamic situations. So far, it has been difficult to make the technology safe and secure in a regulated setting like a hospital. This makes it far more challenging for AI to adapt in contested areas where terrorists or hostile states are actively attempting to block your view.

In such cases, searching for something that the spies have never seen before using AI is a use case for which the technology requires more advancement before it is termed reliable.

Another thing that AI is unable to replicate in human intelligence is the ability to understand information contextually. For instance, Climate change statements from sub-threads of an online debate are presented to an AI system as a set of equally accurate views. This makes them less meaningful than they actually are.

An AI system can also not distinguish between the Ph.D. thesis and the ramblings of a conspiracy theorist unless it has been coded to do so by a data scientist. This presents a chance for the scientist to impose their own biases on the entire data set.

This makes it less effective in spy tasks that require a deeper understanding than AI currently holds.

More importantly, AI is susceptible to security attacks that could compromise the highly confidential status attached to espionage activities. Considering the fact that the quality of an AI system is heavily reliant on the quality of its data, poisoning the training dataset could have adverse effects on the agency and its operations.

Researchers are looking into AI security and safety in the UK to develop strong guidelines for the future. These professionals are working to make sure that elements to prevent adversarial schemes are built into AI design from the beginning.

However, until some of these drawbacks and loopholes are dealt with, AI is still a long mile from substantial application in spy agencies and operations, let alone replacing agents.

Related Articles

• What is Artificial Intelligence? How Does AI Work?
• Meta’s Threads Statistics Continue to Disappoint – Daily Active Users Down 70% From Its Peak
• Craig Wright Will Likely Sue to Claim Copyright on Bitcoin in UK – Here’s What Could Happen If He Somehow Wins