The biggest trends in AI over the past 10 years
Although I enjoy writing about future predictions regarding the world of technology and business, I generally focus on what will happen in the next five years.
Indeed, the companies I advise on data and technology strategy want to know what they need to do now – or very soon – to take advantage of the opportunities of digital transformation.
But it’s also worth taking a longer-term view. I think that in 10 years, the AI that is part of everyday life will be as advanced from today’s AI as today’s Internet is from the Internet of first days.
In his excellent book The coming wave, Mustafa Suleyman notes that each wave of technological change – from the combustion engine to the Internet – has revolutionized society in a shorter period of time than the previous wave. So I don’t think we’ll have to wait 30 or even 20 years before AI is fully ingrained in every aspect of life.
Instead, let’s think 10 years ahead, to 2034. By the middle of the next decade, a lot will have changed, but what will AI look like? Here’s what I think!
Is the AGI quest over?
Artificial general intelligence (AGI) is one of the holy grail goals of AI development.
More specifically, it refers to strong AI who are able to learn to do any job simply by being told what to do rather than by being told how to do it. This contrasts with most Narrow AI of today, which is specialized – designed to learn and improve in a particular task or area of work.
Today’s multimodal baseline models – like GPT-4 – appear to be approaching the capabilities of AGI with their expanded applications. But by 2034, we may have achieved what we consider today to be “real” AGI. This does not necessarily mean that we will produce robots or sentient computers like we have seen in science fiction, but that we will improve current AI systems to work more intelligently and autonomously on tasks daily.
This means that AI will be even more intuitive and effective in helping us at work and at home and will be able to understand and carry out tasks in ways we may never have considered.
However, performing AGI also raises ethical concerns. Although this does not mean that AI will surpass human intelligence (it does singularity, a separate event), we will still need to take steps to ensure that whatever he does is in our best interest. And guardrails will need to be put in place where one’s own ideas of what that means might be different from ours.
Quantum AI
By 2034, it is very possible that the merger of quantum computing and AI will have launched a new era of supercomputing and scientific discovery.
Quantum computing has the potential to significantly speed up certain calculations used in AI algorithms, for example by quickly solving optimization problems that would take much longer with classical computers.
This has already led to progress in drug discovery and material sciences, as well as the efficiency of route planning by delivery companies like DHL. Within 10 years, the accessibility of quantum computing technology will have increased significantly, meaning many more discoveries and efficiencies will likely have been realized.
The emergence of quantum computing is also likely to create significant challenges for society, which could become hot topics by 2024. Specifically, concerns exist that super-powerful, quantum-enhanced computers AI will mean in terms of security and encryption. Some predict that their sheer power means our ability to keep information private could effectively disappear. Hopefully, society will have found solid solutions to these problems by 2034!
AI-powered superhumans
Today we talk about being able to augment our capabilities with AI, but by 2034, has this evolved to the point where we are all basically superhuman?
AI applied to neural interface technology of the type already in use tested in humans could, in theory, improve our cognitive abilities, making us better at learning, remembering information, and even making decisions.
Advances in AI-enhanced prosthetics could give us mechanical limbs or exoskeletons, making us far stronger and more physically capable than we are today. Cameras and optical implants could improve our vision and make old age-related sight decline a thing of the past. And tools that give us insight into human-to-human behaviors and interactions could help increase our empathy and emotional intelligence, allowing us to better understand ourselves.
However, it will be necessary to ensure that these measures do not only benefit the elites. There is already a divide in society between the haves and the have-nots when it comes to technology. Developing ways to ensure that this technology benefits as many people as possible, rather than those who have the most money or live in the most advanced countries, will be a significant challenge for the next decade.
Robots, co-bots and automated friends
Great advances have been made in robotics in recent years, thanks to the application of AI to problems such as balance and moving near humans.
So by 2034, it might seem reasonable to think that mechanical companions will be all around us. Collaborative robots (co-bots) will have left the industrial environments in which they find themselves today to settle in our homes, our offices, our equipment and our public spaces. We will count on them to help us with household chores as well as to accompany and support us and put our lives in their hands if we end up on the operating table.
Robots will easily carry out all types of manual jobs, from deliveries to road repairs and construction. And in parts of the world with aging populations, they will play an important role in providing care and keeping us safe at home. Integrated with natural language technology, they will talk to us, get to know us and develop their own personality. This could lead us to form very different relationships with them than we have with other technologies.
A decade from now, their integration into society could be one of the most important issues facing society. How much autonomy can we safely grant them? And are there ethical lines we should not cross when it comes to treating intelligent entities – even those we ourselves created – as working machines and property?
AI for government and law enforcement
By 2034, AI could be firmly embedded in the infrastructure used to manage public affairs and justice systems. Does this mean we will have an AI Prime Minister who will provide better leadership than a human? Well, depending on your political views, you might say it couldn’t be worse. But I don’t think we’ll be quite ready for AI government in 10 years.
Instead, I think we will have a much deeper integration of smart governance technology into administration. This will streamline processes from managing waste collection routes, issuing permits and licenses, making planning decisions and ensuring services are delivered in a valuable and personalized way for every citizen.
Those involved in legislation and policymaking will use AI to understand what the public wants and set their priorities from there. It will proactively predict issues that may become a concern in the future so that steps can be taken to nip them in the bud. Likewise, police will regularly use predictive technology to anticipate where a crime is likely to occur and even who is likely to commit a crime. These capabilities will be an important part of the conversation in 2034, as we find ourselves increasingly under the surveillance of government and law enforcement technologies, and it becomes increasingly difficult to avoid leaving a digital footprint everywhere you Go on.