WE NEED TO DISTINGUISH between government as in managing a country, which is about the civil service and ministerial departments, and government as in leading a nation, which is all about politics in a democracy, autocracy or dictatorship. (See previous chapter for wider use of AI in surveillance).

AI is already widely used by the civil service in most developed countries, supporting whoever is leading government. Key areas for AI will be developing policy options, modelling the impact of proposed policies, writing policy details and drafting legislation.

And at the end of this book is a proposed 10 Point Plan AI for governments, actions  and policy decisions to take now.

For the next five years, almost all such AI support activities will tend to be on separate systems, but by 2040 many government AI systems will be networked into one all-powerful, all-knowing AI colossus. This is despite inherent risks.

To some extent this is already happening across systems used by security services. So, for example, a security officer can access all of someone’s personal information relating to internet activity (searches, social media, purchases), location, phone logs, family and friends, employment, tax, immigration, banking, driving offences, criminal record, health records – plus other insights generated by AI-linked pattern recognition, predictive analytics. Such an entity will become a primary target for hostile countries to attack (and for ‘friendly’ countries to quietly probe).

Patrick Dixon is a global keynote speaker on AI in Defence, author 18 books

MILITARY SPENDING across the world will accelerate rapidly over the next decade and beyond, because of a growing threat of major wars in addition to existing conflicts. AI will be part of this defence industry boom. Military AI is racing ahead of all other AI innovation, except for AI in cybersecurity, national security, healthcare and workplace automation.

Military forces already spend $5bn a year on AI, for target acquisition, battle simulation, smarter intelligence and better weapon design. The Pentagon alone will invest $2bn in AI in 2025,

some of which will be in cybersecurity and cyberattack capabilities.The first two years since Russia’s full-scale invasion of Ukraine were a ghastly shock to military forces across the world, who had become over-reliant on smart weapons and mobile troops. Half a million fighters trapped in frozen trenches, cowering under artillery bombardment, trapped by landmines, thousands blown apart every month, in static conditions like those in the First World War.

Despite the horrors in Ukraine and the failure of high-tech weapons to achieve rapid gains, AI supremacy will become increasingly important. Indeed, the next stages of the Russian conflict will be dominated by drones and tech which increasingly relies on AI.

Wartime has always driven rapid innovation, and the fastest advances in AI will happen in the heat of fighting, amid the thunder of explosions and the destruction of cities. Advanced AI will detect threats in microseconds and be capable of triggering military conflicts in minutes, beyond the mental capacity of military leaders. AI will therefore end up with huge influence over military strategies and will control most firepower directly or indirectly.

Patrick Dixon is a global keynote speaker on cybersecurity and AI, author of 18 books

THE GREATEST THREAT to our world from AI will be its use by criminal gangs and hostile states. Over the next decade, such attacks will strike hundreds of thousands of companies, destroy wealth, paralyse governments, weaken armed forces, invade privacy on a global scale, knock out hospitals and kill thousands of sick patients. All at very low cost. Most larger-scale attacks make huge amounts of money, from ransoms or rewards from hostile governments.

In my book The Future of (Almost) Everything I warned that combining Big Data with the Internet of Things and the Cloud would create mega-targets for criminals and hostile countries. But next-generation AI will increase these risks a thousand times more. The ultimate target for hostile states will be advanced AI systems of another country.

AI is the greatest opportunity in history to steal money and disrupt countries. 

If cybercrime were a country, it would be the world’s third-largest economy, with 560,000 new types of ‘malware’ released every single day. Most companies have been attacked, while six out of ten are struggling to find people with the right skills to protect themselves.

AI will create new efficiencies for insurance. The benefits of AI to insurers will be huge and will result in lower premiums and greater profit:

◆ More accurate, instant underwriting/quotes based on smarter prediction of risks

◆ Linking of premiums to live data, e.g. miles driven by younger drivers

◆ More accurate grouping of risks for re-insurance with larger companies

◆ Faster and more efficient claims handling

◆ Better fraud detection