The balancing act of using AI as a force for good

AI technology has advanced at vast pace – too fast for mutually agreed legislation, ethics and governance around the globe. There are ethics committees, and region-based treaties, but nothing that defines what is legal or otherwise. That means AI is a potentially powerful tool for criminals and bad actors. Take image manipulation for example. There is nothing to stop someone creating footage of a notable figure being harmed, injured, or killed, and posting it online. If it’s widely believed, it could lead to economic impact or social unrest.

Many of the UK’s enemies have incredible cyber offensive capabilities. They’re not necessarily going to target military installations or government buildings. They’ve found they can disrupt an economy through misinformation, so they’ll hit the public where it hurts. The warfare’s moved from conventional guns, bullets and tanks to just distorting everyday life through false and often inflammatory narratives.

Things aren’t always what they seem online

The seriousness of the threat is captured in a public-facing video from the National Protective Security Authority. Think Before You Link – Glitch highlights that in the online world, things aren’t always what they seem. Simply clicking on an invite to a conference can empower rogue states or criminals. They understand that views of the world can become completely distorted if you get targeted by messaging that twists your personal beliefs a certain way. All algorithms are now AI-driven and we have to be more aware that they’re designed to indoctrinate us more. That video is a milestone. If agencies are putting it out to the public, it’s already a national security level concern.

AI places a finger on the social pulse

However, just as AI can be used by criminals, it can also be used by the good guys. AI-enabled technology has been applied in intelligence for some time, certainly well before its capability became widely recognised. However, while the sector was an early adopter of sorts, it has always taken a cautious approach. The large language models (LLMs) that enable the likes of ChatGPT powered some of the first AI intelligence solutions – and continue to do so. They flag themes and topics that are gaining traction online to analysts. It’s a finger on the social pulse.

Intelligence professionals get bird’s eye view

This is extremely useful to the intelligence sector. With heightened tensions around the world, you get a sense of:

• General sentiment across the population
• Voting behaviours
• False information that is being accepted at face value rather than being challenged

AI-enabled technologies mean huge data sets can be sifted, ingested and processed much faster. This allows intelligence professionals a bird’s eye view, so we see things from a macro perspective. You can tell if certain material, sentiments, or conspiracy theories are influencing people, with a potentially negative outcome for customer organisations.

Treat AI-surfaced intelligence with caution

In Mitie Intelligence, Risk and Advisory Services, we always treat intelligence surfaced through AI as just another source. It requires the same due diligence as a newspaper article, or book or audio snippet. Ideally, it’s corroborated independently by at least three sources. There are occasions when the information can’t be corroborated at all, in which case we use the language of probability to apply appropriate caveats. That means reporting with terms like ‘almost certain’, ‘realistic possibility’ and ‘highly likely’. The approach is aligned to the Professional Head of Intelligence Assessment (PHIA) standard.

We’re entering 1984

For anyone familiar with George Orwell’s dystopian novel, 1984, it reads like a manuscript of the future we’re entering. Cameras everywhere, people being monitored, Big Brother watching. AI has the power of facial recognition, which is incredible and a brilliant solution for preventing or reducing crime. But as responsible intelligence professionals, we’re very aware of what it can do versus how it should be used.

Use AI as force for good

Ultimately, using AI for intelligence purposes is a balancing act. It’s incredible capabilities must be weighed alongside privacy legislation and ethics. Intelligence teams should look to what’s ahead with excitement and caution. In the wrong hands, there is a lot that AI could do to disrupt the world. We just need to make sure those on the right side of the law are empowered to use it as a force for good, through strict governance, ethics and care.