We’ve all experienced the unprecedented increase in discourse around AI over the past year. This has ranged from publications about the humble history of AI to dramatic speculations that the technology could evolve us from skin-bag ground walkers into spacefaring cyborgs within the next century.

As much as I’d love to bounce around on the Moon when I retire, I think the possibility is a bit further away than my state pension – since I’ve got another 30 years to go. However, from looking at what’s been going on in the past couple of years, we can speculate about where we might be in the next decade with the help of AI.


To make this a bit easier to digest, I’ve split it into three sections:

  • A Recent AI History
  • Present Day Trends
  • What’s next for AI?


A Recent AI History

To understand current trends in AI, we need the context of where it all started.

So, here’s a handy summary in the form of an infographic:


Present Day Trends

With all the recent leaps and bounds in AI technology, there are lots of trends that have emerged over the past year. 

Uneven AI Adoption

The sheer speed of AI’s growth and adoption sets it apart from any technology that’s come before it. 

  • The number of UK AI companies has increased by 688% over the last 10 years. 
  • 68% of large companies, 33% of medium-sized companies, and 15% of small companies have incorporated at least one AI technology.
  • The UK AI market is worth more than £16.9 billion, according to the US International Trade Administration, and is expected to grow to £803.7 billion by 2035. 

However, this pattern isn’t globally consistent. 

While the UK and the US have been fast to embrace and enable AI’s applications across industries, other countries still need to craft a market in their economies for the technology to grow. Otherwise, it’s down to private companies, which could leave AI growth stunted in some areas.

Opposing AI Opinions

Current discourse shows a sharp divide on whether accessible AI language and generative models are a useful tool, a hindrance, or just a plain nuisance.

Concerns range from generative AI content tools making us lazy and killing our creativity, to creating issues around plagiarism and cheating in an academic setting. The worries have triggered extreme reactions from educational institutions desperate to preserve their prestige. Since its introduction in November 2022, eight out of the 24 Russell Group Universities have formally banned the use of ChatGPT and other AI language models.

There are, of course, limits to generative AI’s usefulness. In an educational setting, the input of the human is infinitely more valuable, as the end goal is self-betterment. However, when it comes to the workplace, where efficiency is of the essence, generative AI tools have a valid and productive purpose. Employees across industries are keenly embracing them for help with repetitive, administrative tasks.

In fact, data from Accenture recently found that 53% of UK full-time and part-time employees in white-collar occupations use generative AI tools at work at least once a month.

We can only speculate on whether cautious and enthusiastic views towards AI will merge, reconcile, or remain at odds in the future. As the capabilities of the technology evolve, no doubt new uses and debates will continue to emerge. Will you be an eager adopter or a cautious naysayer?

The Job Market

Currently, no one knows the long-term impact of AI on the economy and jobs, but predictions are coming in thick and fast from a range of organisations and sources – with varying levels of optimism.

Some claim there will be a sudden boom in ‘AI jobs’, followed by a sudden fall as generative AI models become more self-sustaining, requiring far less training from external sources.

Others fear that AI will steal jobs at all levels, from administrative roles to creatives. Even the UK government estimates around 7% of existing UK jobs could be displaced over the next five years, rising to around 18% after 10 years and nearly 30% after 20 years – equivalent to around 2.2 million jobs.

Despite this, still others claim there is nothing to worry about, as AI still relies on data-fed models, which require humans to create, train and monitor. As much as it seems like it, AI is not autonomous or, arguably, even capable of thinking for itself at the moment.

All we can say for sure is that AI has already had a massive positive impact on the economy in its short life so far, contributing £3.7 billion in 2022. Britain especially has reaped some exciting rewards by embracing AI innovation – we have twice the number of AI-based companies than any European nation. What’s more, job listings across all sectors that mention AI, advertise salaries 20% higher than those that don’t.

While I can’t predict the future, at the moment it looks like AI is having a more positive impact on the job market than negative.


Multimodal AI incorporates the interplay of different representational modes such as images, videos, spoken and written words and numbers. It is capable of more accurate determinations, more insightful conclusions, and more precise predictions about real-world problems. It is essentially the next evolution in generative AI’s capabilities.

While multimodal AI offers the most promising features of AI yet, it also poses the most threats to the way we currently work. Its ability to work with different modalities means it can be applied to ‘translate’ content – whether that’s from text to image, image to video, text to audio, or virtually any combination of the above.

These types of AIs cannot only take on more creative tasks but more data-led tasks too. Imagine being able to upload a set of spreadsheets and have an AI model read and analyse them for insights, then create a custom PowerPoint summarising those insights too.


Our Predictions of Future Trends

We don’t have a crystal ball, but we do have data to inform us of what the future could hold for AI.

Training Costs Decrease

Over the past year alone, the costs associated with training deep learning AI models have plummeted as the technology has advanced.

According to the ARK Invest Big Ideas 2023 report, training costs of a large language model similar to GPT-3 level performance have plummeted from $4.6 million in 2020 to $450,000 in 2022, a decline of 70% per year. 

According to Unite.AI, factors contributing to the cost decreases include:


  • Hardware: AI-relative compute unit (RCU) production costs, i.e., costs associated with AI training hardware, should decrease by 57% annually, leading to a 70% reduction in AI training costs by 2030.
  • Software: AI software training costs have the potential to be lowered by 47% annually through increased efficiency and scalability. Software frameworks like TensorFlow and PyTorch enable developers to train complex deep learning models on distributed systems with high performance, saving time and resources.
  • The Cloud: Cloud-based AI training reduces costs by providing scalable computing resources on demand. With pay-as-you-go models, businesses only pay for their computing resources. Cloud providers also offer pre-built AI services that accelerate AI training.

Matching Human Performance

According to McKinsey,

“Generative AI will perform at a median level of human performance by the end of this decade. Its performance will compete with the top 25 percent of people completing any and all of these tasks before 2040. In some cases, that’s 40 years faster than experts previously thought.”

However, human-level performance doesn’t necessarily mean humans are no longer needed in the world of work. Instead, it means that we can hopefully get menial, boring and repetitive tasks performed to a human standard (or higher) without having to do them ourselves. That means more time to think big and get creative with our roles.


AI in Education

Offline tests done with pen and paper involve significant admin and logistics – often on a national scale. That makes them time and resource-intensive, especially with the growing number of students and educational institutions. 

In contrast, online exams offer a more efficient and streamlined alternative but have been typically harder to monitor for cheating.

AI is now playing a transformative role in the education industry by automating administrative tasks, such as grading assignments and managing attendance. In the context of online exams, AI contributes to convenience and security, revolutionising the way assessments are conducted and providing valuable feedback to educators through the analysis of student test results. 

However, as we said earlier, the opinion on AI in the current education system is pretty low. Especially at a university level, where students are encouraged to form their own opinions and theories. For AI to have a real impact on education, there will need to be a revolutionary shift in thinking around the technology.

Short-term Workplace Momentum

In the short term of around three to five years, colleagues will encourage each other to use and adapt AI, sharing knowledge and raising AI literacy amongst the general working population. 

Now that Gen Z and Gen Alpha are entering the workplace, this is likely to accelerate further. 

While some tasks are replaced with AI, most jobs will be reskilled to be augmented with AI, not lost altogether.

My Predictions

Now you’ve heard the official narrative, here’s my take on the next big developments in AI.

Video games

Video games, especially those using VR and AR, will be revolutionised as AI becomes more advanced. Game designers currently have to design levels and game environments manually. In the next few years, AI could easily design high-quality in-game environments, speeding up production.

Anti-AI content

AI has seen plenty of resistance from people in creative industries. I don’t see that resistance going anywhere. I believe all-human content may become an even more appreciated art form, the rarer it becomes. While this may be years in the future, I believe there will be a small portion of the internet dedicated to content that has no AI involvement at all – not even primitive forms like spell-check.

Academia’s own AI

While the academic world has so far proven wary of AI, and for good reason considering its loose grasp on facts and enabling of cheating, I don’t think it will be able to resist the technological tides much longer. AI will change the nature of studying over the next few decades, There will soon be specialised AIs made for studying. Whether these are going to be accessible, or privatised for profit like in-person tutors, is another discussion.


Final Thoughts

  1. Artificial intelligence is on a dynamic journey. Lots of changes have already taken place, and tons more are on the horizon. From its history to present-day adoption around the globe, the discourse surrounding AI’s development is marked by opposing opinions regarding its impact on creativity, education and the job market.
  2. The rise of multimodal capabilities and the decreasing costs of training deep learning models present both promising opportunities and potential challenges. Predictions made from all corners of the tech world highlight a short-term momentum in workplace AI adoption, and transformative roles in sectors like education and marketing.
  3. No one has a crystal ball granting clairvoyance on what’s to come in the AI world, but we do know that it’s going to impact lots of different areas.  Make sure you’re using the tech as a tool to augment your work, rather than pushing against the rising AI tide.