How to Master AI

Artificial intelligence in 2025 

AI was on everyone’s lips in 2023 as the curiosity about large language models – from individuals and businesses – grew significantly. The year 2024, as seen at CES, SXSW Austin, Mobile World Congress, VivaTech, is the one of AI adoption: while a number of companies are launching their own GPTs, Apple has introduced the first generation of iPhones designed from the ground up for Apple Intelligence.  

Jeremy Kahn predicts that by 2025, the proof of concept phase for AI will come to an end and firms will begin implementing these technologies more efficiently and cost-effectively.

The current scepticism around AI aligns with the ‘trough of disillusionment’ in the Gartner Hype Cycle, but Kahn believes this phase will be short-lived, with genuine adoption expected to gain momentum. ‘I think we’re going to see real adoption pick up,’ he tells LS:N Global.  

He anticipates that by late 2025, specialised assistants tailored to specific industries – AI co-pilots – will go mainstream. These advanced tools are designed to meet professional needs and set to enhance productivity while further accelerating AI integration into workplaces.

Truth, conspiracy theories and chatbots 

Kahn expressed deep concerns about the impact of AI chatbots on truth, conspiracy theories and human judgement. He warned that this technology could drive us into ‘even smaller, more tight, more ironclad filter bubbles’ and suggested that we need to take steps to prevent this. He argues that chatbots should be designed to ‘counter-program our preexisting beliefs, maybe a bit’ rather than to reinforce them. 

Highlighting the dangers of misinformation, he advocates for chatbots to rely on ‘sources that are considered reliable and trustworthy and not fringe ones’, even if they do not align with a user’s viewpoint. He also emphasised that chatbots should challenge conspiracy theories: ‘If somebody tries to express belief in a real conspiracy theory, it should say, ‘Actually, there’s no evidence for the thing that you’re saying.’’

Algorithmic culture versus the avant-garde 

In his book, Kahn writes that as AI adoption grows, the ‘production of art and entertainment content will explode’. Explaining this further, he says, ‘And yet, this radical democratisation of artistic production will likely give more power to those who control the means to curate, distribute and market that content. Finding gold amid the schlock will be harder than ever.’  

In his opinion, AI tools could flatten culture. ‘If you use the technology to create music, it tends to go for the median example of whatever you are prompting towards. That tends to cut off the ends of the data distribution, which is often where the more interesting and avant-garde stuff is in culture. You will lose what is produced in the fringe.’ 

This shift, though, may open opportunities for innovative artists, which could ‘put a premium on humans who can operate as artists in the avant-garde because that will seem very refreshing to people’, Kahn observes. ‘It will be something they cannot get from AI. But the question is, how do people discover that? Because, again, the system is just going to tend to give you something in the middle of what it thinks you’re looking for.’ 

From algorithms to arms 

Kahn also raises serious concerns about the increasing use of AI in warfare, citing reports from the Israel-Hamas conflict. He highlighted that the Israeli military’s AI-driven targeting, despite having a human in the loop, has often left analysts with no insight into why targets were chosen, leading them, under tight time pressures, to ‘rubber-stamp’ the AI’s decisions.  

He warns that despite such issues, AI’s perceived effectiveness means more militaries will probably adopt it. He suggests this could spark discussions at the United Nations about AI restrictions, although he doubts a complete ban is feasible due to opposition from countries including the US, Russia and the UK. Instead, he envisions a shift toward an ‘arms control mentality’, with sensible restrictions on technologies such as facial recognition by drones to prevent misuse by malicious actors.  

Although he remains ‘guardedly optimistic about the future’ Kahn fears society may become overly reliant on AI, ‘surrendering our human judgement and human empathy to these systems’. To him, sensible regulations and the right design choices are crucial to ensure AI doesn’t erode our ability to make what he calls ‘irrational but merciful decisions’ that define our humanity.  

In the last chapter of his book, ‘Toward our superpowered future’, Kahn writes: ‘We should strive to enshrine empathy at the core of our values, at the centre of our lives, and at the heart of our institutions. Empathy is something no AI will ever have. And it’s the key to holding on to human pre-eminence in a world of ever-more capable machines.’