Time to think – that is what we always seek from new technologies. If machines do the boring bits, we can concentrate on the important stuff, such as thinking.
However, thinking and other creative endeavours are precisely what we are now outsourcing to artificial intelligence (AI). It can read, write, draw and compose music. Soon it will resolve legal cases, diagnose patients, encode systems and design buildings. If AI takes over the traditional areas of human activity, what will remain for us to do?
We humans naturally strive to make our lives easier, even when it is not good for us. Our muscles have already atrophied from lack of daily physical activity. Will the same happen to our brain?
The situation is not as bleak as that, believes Postdoctoral Researcher of Practical Philosophy Pii Telakivi. The brain changes according to the way it is used. When we have developed new tools, our memory and numeracy skills have not declined, but changed.
“According to Plato, Socrates complained of writing destroying memory. This didn’t happen. In fact, writing supports thinking and memory. Likewise, AI applications can help thinking.”
And the only reason we know of Socrates’s worries is that Plato wrote them down.
Basic skills still crucial
Humans are inclined to avoid difficulties. Often this makes sense.
“But you can’t always tell which apparently dull activities are key for your wellbeing and learning. It will be worrying if we don’t develop our basic thinking skills as a result of excessive reliance on AI,” notes Telakivi.
What actually is thinking? It involves at least remembering, learning, deciding, solving problems, using language, emoting and observing.
“AI use can affect memory and learning both positively and negatively. It can be very helpful for observing, resolving problems and using language.”
But AI is incapable of critical thinking. Judgement, or the ability to evaluate the validity of information, is becoming increasingly important for humans. It is possible that AI will in fact make people more critical than before, Telakivi suggests.
Fixing neural connections
An untravelled forest path will be reclaimed by the wilderness. Similarly, neural connections in the brain will either weaken or strengthen depending on how they are used. We have the capacity to create new connections and lose old ones.
“Without practice, brain function and the ability to concentrate won’t reach their full potential,” says Mona Moisala, Doctor of Philosophy (Psychology).
Taxi drivers have been found to have an increased volume of grey matter in brain areas linked to spatial memory and navigation. If you always rely on map applications, these areas will be weaker.
The younger the person, the more important it is that their brain receives appropriate stimuli during each sensitive period. This may occasionally require tedious mental tasks as well. A general upper secondary school student can outsource an essay to ChatGPT, but by doing so, they are developing ChatGPT, not their own brain.
AI requires an expert able to use and supervise it as well as determine whether the outcome is correct. How will such experts develop if we rely on AI in our education and thus miss out on the stage that refines our expertise?
As for older adults, they need stimuli to maintain neural connections in the brain. This is beneficial if they later develop dementia.
Brain development requires practice
We can give AI tasks that the human brain is unable to complete. To save time, we can also give AI tasks that our brain is able to complete.
“We’ve had this rosy scenario of technology freeing up time. So far, this hasn’t happened, as we’re continuously taking on new tasks. I’m afraid AI will only accelerate this development, leading to an ever-increasing burden on people,” says Moisala.
If time really is freed up, what will we do with it? One of the things vying for our attention is social media.
“In the early stages of social media, people didn’t understand how algorithms affected them.”
It is only now that we have discovered how bad they are for our mental health and ability to concentrate. But we have opened Pandora’s box, and there is no turning back. This means it is important to consider the nature and effects of AI apps in advance.
Moisala is also concerned that when thinking is outsourced, the effort that develops the brain is left out. It is precisely this effort that has made us human. Without exertion, the network of connections between brain cells can become impoverished.
A bigger change than the internet
Over the past few years, AI systems have developed to the point where they are now capable of interpreting natural language and giving answers. This opens opportunities for both new services and scams.
The situation may even surpass the advent of the internet, says Professor of Computer Science Sasu Tarkoma.
He believes that within just five years, artificial intelligence systems will revolutionize industry and society. They may also lead to significant disparities in skills and economic status.
“We can’t rely on AI systems to function correctly or be used correctly. We need mechanisms and regulations to ensure this. And we must train people to use the systems.”
Deteriorating confidence in others
Unaided, we find it difficult, if not impossible, to distinguish AI-generated photos and videos from authentic ones. We must always be on the lookout for forgeries. Then again, embarrassing genuine videos are easy to brush off as fakes.
Earlier this year, a major theft took place in Hong Kong with the help of AI. In a conference call, a fraudster posed as a bank worker’s supervisor and colleagues, convincing the worker to pay out over €20 million.
When we can no longer trust our own eyes and ears, our trust in others may deteriorate. Tarkoma expects the emergence of a new industry focused on confirming the veracity of information. The importance of verification and authentication will grow, driving up costs.
“What can we trust when quantum computers break encryptions and artificial intelligence falsifies everything? Maybe our society wasn’t ready for so sophisticated AI with the ability to scam on demand.”
For human needs
Sasu Tarkoma says that we should strive to develop human-centred AI based on human needs and with humans responsible for the whole process.
Pii Telakivi agrees: AI solutions should be used to solve existing problems. But she decries the priority usually given to the market perspective.
“Tech companies bring apps to the market, and all of us adapt to them, regardless of actual need. We begin to comply with their requirements, and soon no one remembers where it all began and why we ended up like this. It’s what happened with social media.”
Technology has changed people ever since the Industrial Revolution. A major leap took place with the arrival of smartphones.
“Checking your social media accounts once a day on your laptop is different to always carrying them in your pocket or hand,” Telakivi points out.
Our brain needs screen-free time. We should engage in offline activities, such as going out for a walk to calm our mind.
Keeping an eye on technology
In the past couple of years, artificial intelligence has started to impact jobs in creative fields. Previously, automation swept through industry and reduced factory work. Before that, industrial production did the same to craftsmanship.
“Freely created art is unlikely to be lost. It is so fundamentally tied to the meaning the artist is conveying. But work associated with the creation of commercial content will change,” says Telakivi.
It is possible that many future jobs will involve giving prompts to AI and cleaning up its messes. It will not necessarily feel as meaningful as seeing your own distinctive touch and controlling the process from beginning to end.
In industry, workers have long been responsible for monitoring technology. We know less about how humans and AI collaborate in knowledge-intensive work, with humans identifying faults.
“It may require several humans and several AI systems. It’s possible organisations will need whole teams to ensure their operations are not disrupted,” says Tarkoma.
And will it be possible to find consumers for all AI-generated content? People are reading less and less, and when the number of texts available online multiplies, they will perhaps not be read by anyone other than AI itself.
A whole new level
Yet, AI outperforms humans in tasks including observation, data collection and calculation. In these areas, AI has the potential to take us to a new level altogether. Human trials on implants allowing us to interact with the digital world are already underway, and humanoid robots are being tested.
Tarkoma believes AI offers great opportunities for research as well. Automated laboratories and experiments can deliver breakthroughs, for example, in medicine and pharmacy. This may lead to a longer lifespan and better functional capacity for us humans.
“On the other hand, the outcomes may not necessarily be distributed democratically,” he says.
Similarly to most aids facilitating people’s lives, AI requires a tremendous amount of energy. This may encourage people to develop clean energy – or increase demand for fossil fuels.
The changing face of humanity
Although AI may be an effective tool for decision-makers, the risk is that they will begin to rely too much on it, losing confidence in their own ability to make decisions.
“Where’s the line between feeling you’re in control and being manipulated? When you use a device, do you feel you control it, or vice versa?” Telakivi asks.
“We’ll become better at telling AI what to do, but we may also lose something. This is an important issue to investigate,” says Tarkoma.
Is some essential aspect of humanity disappearing? If I no longer think myself, does it mean I will therefore cease to be?
“Being and thinking will not cease, but they may change, which means humanity too will change,” says Telakivi.
Figuring things out with our own minds also gives life meaning. Let’s not give up the joy of understanding, whatever we may become.
The article was published in Yliopisto magazine 7/2024 in Finnish.
It matters where in the world artificial intelligence is located and where it has been developed. AI solutions differ depending on the cultures and values of the countries where they are built.
At present, AI development is driven by major corporations which can decide with whom to collaborate and can terminate agreements at their convenience.
The seven biggest tech companies all hail from the United States. Their combined market value is the same as the combined gross domestic product of Germany, France, the UK and Italy.
“We’ve already lost the digital infrastructure. Most digital services are bought from outside the EU. Going forward, this will probably happen to AI services too. The EU hasn’t been very good at creating a new business model for the tech sector,” says Professor Sasu Tarkoma.
This means a large transfer of income from Europe to the United States. Even though AI is also being developed in Finland and elsewhere in Europe, catching up with the Americans will be difficult. AI requires investments of up to hundreds of billions of euros.
Finland has the advantage of hosting one of the world’s largest public supercomputers, LUMI, in Kajaani. LUMI is powered by 10,000 AI GPUs. For comparison, Meta recently announced it will operate 600,000 AI GPUs.
A key element of geopolitical competition, the primary production of these electronic circuits, is dominated by South Korea and Taiwan. It demonstrates that, if you have special expertise, you do not have to be a superpower to succeed.
In addition, Tarkoma pins his hopes on open-source AI, launched with great momentum. It will foster transparency and collaborative development.
“It is important that we are involved in developing technology based on the needs of society and for society, and that we are able to utilize the lessons and results also in business.”