22% of 18- to 21-year-olds have sought emotional support from generative AI.oatawa/Shutterstock
Some of us now use generative artificial intelligence (gen AI) to talk through difficult feelings. We may ask a chatbot to write an email or plan a holiday, but we also tell it that we have had a rotten day or feel anxious about a presentation.
A 2025 US survey of 1,058 people aged 12 to 21 found that 13% had sought advice from gen AI when feeling sad, angry or nervous. Among those aged 18 to 21, the figure was 22%. These findings suggest that turning to gen AI for emotional or mental health support is already relatively common among young people.
So is gen AI a poor substitute for a friend? Human-written replies may seem the obvious source of greater support. However, in some experiments, including our own, people have rated gen AI responses more highly.
Our team at the universities of Manchester and Durham conducted five experiments comparing emotional-support messages written by people with those produced by large language models (LLMs), the systems behind chatbots such as ChatGPT.
In one experiment, 390 participants imagined situations involving anger, sadness or fear. They then read either a human-written or gen AI response without being told its source. For the anger and fear scenarios, participants rated the gen AI messages as more emotionally supportive. There was no reliable difference for sadness.
Gen AI messages also improved some of the emotions participants reported, although the pattern varied. In the fear scenario, for example, they increased calm more than human messages but did not produce a greater reduction in fear itself.
Our findings fit a recent review of 23 studies. Overall, people tended to rate gen AI messages as more empathic, meaning that they appeared understanding and caring, than human-written ones. Researchers call this the “AI advantage”.
One possible reason is consistency. Gen AI can produce a structured response on demand, while people may feel tired or unsure what to say. It also tends to include a concrete suggestion that the recipient can follow.
Our experiments tested possible explanations for this advantage. In one experiment, explicitly acknowledging and validating the recipient’s feelings did not account for the greater emotional improvement associated with gen AI.
Two further experiments found stronger evidence for what we called “actionable support”: specific and realistic suggestions. Messages containing this practical help were rated as more comforting and improved some emotional responses, regardless of whether they came from gen AI or a person. When human messages offered comparable help, they were judged just as supportive in these comparisons.
Practical support also requires restraint. A 2024 study found that excessive suggestions could be less effective at making people feel heard. Before offering solutions, it may therefore help to ask whether advice is what the person wants. If it is, one manageable step may be more useful than a list of remedies.
Why do we still prefer people?
The source attributed to a message changes how we receive it. The same 2024 study found that gen AI responses made people feel more heard than human-written ones, but this benefit declined when recipients believed the message came from AI.
Across nine further studies involving 6,282 people, the same gen AI responses were rated as more empathic and supportive when described as human-written rather than attributed to AI. Participants also preferred human interaction when seeking emotional engagement, even when choosing a person meant waiting longer.
Research has yet to establish exactly why. One possibility is that a human response signals someone’s willingness to spend time and emotional effort. Support from somebody close also arrives within an existing relationship.
Our research with romantic couples offers indirect evidence of the importance of the recipient’s perception. We asked one partner what they did to help the other manage difficult emotions, and asked the recipient what they believed their partner had done. Both people also rated their relationship.
The recipient’s perception of their partner’s efforts was more consistently associated with both partners’ ratings of their relationship than the helper’s own account. A separate study found that valuing and receptive listening were associated with greater relationship satisfaction.
These studies identified associations. They did not establish that particular forms of support caused better relationships or explain why people prefer human support. They do, however, show what comparisons between isolated messages leave out. Comfort from somebody close comes with personal knowledge and the possibility of continued involvement. Gen AI can suggest going for a walk, but it cannot come along.
The wider research also has important limits. Much of it tested brief or one-off exchanges, often involving imagined everyday situations, and measured immediate reactions. It cannot tell us what happens when somebody relies on a chatbot over months or uses one during a mental health crisis.
The human messages were generally written by strangers or developed as representative examples. The experiments therefore did not compare gen AI with support from a friend who knew the recipient and understood the wider circumstances.
The practical lesson is that people can learn from what gen AI does effectively. When somebody is upset, ask whether they want advice. If they do, offer a specific next step that feels manageable.
Gen AI may help us find useful words and possible responses. Human relationships add personal knowledge and the ability to remain involved after the message has been sent. Used carefully, gen AI could help us become better at supporting one another.
The authors do not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.
Pelicans have resided in St James’s Park, London for 360 years. But the current flock have been a summer sensation due to the first hatching of chicks in the park’s history.
Pelican breeding in zoos is demanding for these birds, because the right environmental and social conditions need to come together. Rather than simply being comfortable, the birds need to encounter and respond to cues that stimulate courtship, nesting and reproduction.
This unusual nesting at St James’s Park provides an interesting example of how responding to biologically relevant stressors is a positive part of an animal’s life. Multiple factors are likely to have contributed to the pelican nesting. The heatwave, the flock’s dynamics and the compatibility of the individual birds seem to have created the right conditions for the first time in centuries.
In my recent research, I evaluated the importance of positive stress to zoo and aquarium animal welfare and found it is crucial for the wellbeing of captive animals.
Researchers call the beneficial response to a stressor, “eustress” – eu being the Greek prefix for good.
Courtship, reproduction and chick-rearing are all natural stressors. Priorities change, energetic and physical needs alter, and adults spend more time on offspring. This stress, however, is not necessarily bad. For example it can provide learning opportunities and enhance the animals’ behavioural repertoire.
As we learn more about what good animal welfare is, it is becoming clear that meaningful challenges which reflect an animal’s natural ecology help them have more autonomy over how they live. Eustress also helps animals cope better with averse situations.
Zoos do not need dramatic changes to provide these experiences. Small challenges can be built into everyday care: food can be hidden in different locations, or new scents and natural materials introduced. Animals can encounter new ground materials and surfaces (such as soil, bark and sand), changes to ropes or branches, or vegetation that naturally changes with the seasons. Enrichment can also encourage animals to explore different parts of their existing enclosure or use familiar resources in new ways.
Variation is key. If food always appears in the same place, at the same time and in the same way, there is little need for an animal to make decisions or adapt.
Many zoos already provide these opportunities. The next step is to think more deliberately about the challenge itself: is it meaningful, does the animal have a choice about engaging, and can it successfully overcome it? The aim is not to make life unnecessarily difficult but for animals encounter novelty or problems they can solve.
Rather than simply giving an animal what it needs to survive, we should provide opportunities for the animal to use their skills and make decisions.
What might this look like at your local zoo then? Meerkats (Suricata suricatta) are one example. Wild meerkats do not encounter their food neatly presented in a bowl. They need to sniff, dig and investigate their surroundings, while members of the mob keeps watch for danger.
Finding food is also a social challenge for these animals. Young meerkats learn about unfamiliar foods by observing and interacting with more experienced adults. Older meerkats help youngsters acquire important food handling skills.
Zoos can recreate elements of these challenges by scattering or hiding food, providing a sandy floor that encourages digging and creating elevated lookout points. Likewise, zoos can encourage social opportunities by providing complex feeding experiences where animals forage together. A 2021 study found that zoo meerkats retain the sentinel behaviour seen in their wild counterparts, suggesting that opportunities to perform these sort of activity remains important even when predators are absent.
The same principles apply underwater. Common clownfish (Amphiprion ocellaris) also encounter environmental and social challenges. On coral reefs, clownfish forage in a complex three-dimensional environment and depend on a host anemone for protection. They live in a strict social hierarchy. The largest fish is the dominant female, followed by the breeding male and then smaller, non-breeding fish. This social group is known as as a “queue”, as subordinate fish can move up through the hierarchy. In the aquarium, recreating this social group by housing a shoal of clownfish around the correct host anemone species allows for more choice and for fish to express a richer behavioural repertoire.
This brings us back to St James’s Park’s pelicans. Successful breeding has introduced perhaps the most natural challenge of all. Finding a partner and rearing young to maturity requires animals to change how they spend their time and respond to altered demands. When animals work for a reward, scientists call this contrafreeloading.
However, zoo animals can experience negative stress when a stressor is prolonged, overwhelming, unnatural or beyond their control. Examples can be excessive noise, incorrect climate conditions or being unable to retreat from something they find threatening.
The lesson for zoos is a good life should not necessarily be an entirely effortless one. Welfare is enhanced when animals can express behaviour their evolutionary history has equipped them to perform. Eustress is not giving every zoo animal the same kind of “challenging enrichment”. The challenge needs to make biological sense for that particular species.
Did you click on this article out of your own free will? Or was it predetermined by the cosmic evolution of particles, unfolding in an unbroken chain reaction set in motion by the big bang? Perhaps you think random quantum processes or unconscious brain activity were responsible? Whether you are philosophically inclined or not, the topic of free will is something most of us have pondered at some stage.
Research has shown that people tend to believe in free will, at least in some form. A more important question, perhaps, is whether it matters. This is the exact conundrum that researchers in psychology and experimental philosophy have been trying to answer recently. What would ultimately happen if people gave up on the idea of free will?
It may seem like a dangerous idea; something that could result in chaos and depression. And indeed, research has largely found that when you manipulate people to doubt free will, bad things happen – including an increase in behaviour such as cheating and aggression. This suggests that even if free will doesn’t exist, as many scientists and philosophers assume, it might be better for us to pretend that it does. But newer research by me and others suggests that doubting the existence of free will may not have such bad consequences, after all.
Philosophers have argued for centuries about what free will is and whether we have it. Positions vary widely, and it is impossible to give the subject full coverage here. At one end of the debate are libertarians, who contend that at least some of our choices and actions are not causally determined. At the other end are hard determinists who argue that every thought, action or event is a result of past events and the laws of nature.
When people talk about a disbelief in free will, determinism is usually the position that springs to mind. Yet, determinism is one of many positions within a family of views that dismiss the notion of free will. A position that has emerged more recently is called free will scepticism. While this position rejects the idea that humans possess genuine free will (but not necessarily agency), only some sceptics reject it because of determinism, while others argue that free will is impossible in an indeterministic or random universe.
But whichever tribe you identify with, what are the actual consequences of your beliefs? Before looking at some of the issues with recent experiments, let’s first explore how researchers began manipulating beliefs in the first place.
The Insights section is committed to high-quality longform journalism. Our editors work with academics from many different backgrounds who are tackling a wide range of societal and scientific challenges.
The benefits of believing in free will
The first piece of research to experimentally manipulate free will beliefs was conducted by Kathleen Vohs and Jonathan Schooler in 2008. This work, which is now a seminal paper in the area, comprised two experiments. In the first experiment, the researchers asked participants to read text taken from Francis Crick’s classical book The Astonishing Hypothesis. Participants were allocated to read either an anti-free will version or a neutral one.
In the group that was manipulated to not believe in free will, the text stated that free will was illusionary and behaviour was driven by unconscious brain activity. For example, the text included sentences such as “everything people are and do is the product of simple, physical processes in their brains” and “although we appear to have free will, in fact, our choices have already been predetermined for us and we cannot change that.” In the control condition, the text did not mention free will and instead discussed research on consciousness.
After reading the text, all participants completed a mental-arithmetic task on a computer. They were told that a “programming glitch” meant that the solutions/answers to the task that they were meant to solve themselves would be visible. Participants were therefore encouraged to press the spacebar, which would ensure the solutions remained hidden. Therefore, whether or not participants decided to keep the answers hidden was used to indicate cheating.
The researchers speculated that those reading the anti-free will text would press the spacebar less so than the control condition. And this is what they found. To confirm these findings, they ran a second experiment with some slight methodological changes. This second experiment essentially found the same thing: participants made to disbelieve in free will cheat more.
This paper led to a proliferation of studies seeking to examine other outcomes associated with free will beliefs. Roy Baumeister and colleagues investigated the link between the belief in free will and aggressive behaviour. After a free will belief manipulation, participants completed a cooking task and were told that the individual they were preparing food for disliked spicy food. Therefore, the amount of hot sauce administered by the participant served as the measure of aggression. Consistent with the researchers’ prediction, participants whose belief in free will had been weakened slathered on more hot sauce. The researchers therefore concluded that belief in free will is related to aggression. That said, they only looked at the direct consequences of manipulating people doubt free will rather than capturing the exact thought process of why they used more hot sauce.
The consequences of this programme of work were clear: if we make people disbelieve in free will, negative outcomes will follow. Consequently, even if we believe free will is an illusion, people should be shielded from this information as a society. However, there are some issues with these manipulation studies and the catastrophic conclusions derived from this line of research.
Are these conclusions premature?
An important issue concerns the beliefs that these manipulations are targeting. As mentioned earlier, many arguments against free will exist, and each argument aligns itself to a different philosophical position. The texts given to participants typically make claims related to the inevitability of events, the nature of science and the role of consciousness.
They are an amalgamation of philosophical positions that have their own ramifications. One idea that is commonly used in manipulations is the suggestion that people’s thoughts, desires, and wills are ineffectual – that the brain prepares for action before we consciously decide to do something. So when you reach for that chocolate bar at the end of a hard day, it’s not a result of conscious choice – unconscious brain processes are entirely to blame.
But this is not the only way think about a lack of free will; and it’s not something a sceptic would endorse. Many sceptics believe that thoughts and desires do influence how people behave but, importantly, the cognitive processes that underlie them (including genetics and past experiences) are either determined or due to chance. This suggests your desire was at play when you ate that chocolate bar – you can have agency without having full, God-like control deserving of ultimate moral responsibility. In this view, free will is more of a metaphysical concept whereas agency is involved in day-to-day events.
This is an important but subtle difference. Telling people that their thoughts have no say in what they do might lead to feelings of demotivation, disillusionment and alienation. It may even lead them to immoral behaviour. But these consequences are likely a result of far-fetching, dramatised arguments that pose direct threats to agency, and not a general, sceptic approach.
Suppose the manipulations do weaken belief in free will, but only according to a sceptical position that maintains some belief in agency. Will it still lead to dramatic, abhorrent outcomes? Recall that the experiments with the mathematical task, with cheating behaviour represented by the number of keypresses that did or did not occur.
The most that can be concluded from this experiment is that weakening belief in free will can influence cheating in a fictitious mental arithmetic task. The study did not find a relationship between belief in free will and other, more socially relevant, ways of deceptiveness, such as fraud, infidelity, academic dishonesty and so forth. To assume people cheating in a controlled laboratory experiment, which has no real-life consequences, would also commit these other forms of cheating appears far-fetched.
Similarly, the cooking study showed that beliefs about free will can affect how much hot sauce participants allocate in a laboratory setting. It is possible that the same manipulation could also increase verbal hostility or confrontation, and perhaps even aggression or threatening conduct. However, these conclusions remain completely speculative – we just don’t know.
For argument’s sake, let’s assume that they do lead to bad things in real life. We may still question just how meaningful the findings are. Researchers talk a lot about the “effect size” of a finding, which basically tells us the magnitude or size of the effect. A large effect size indicates a finding has practical significance, whereas a small effect indicates limited practical application. Researchers obviously strive for the former.
To get a sense of the effect size in free will belief research, a 2022 paper reviewed studies administering manipulations to free will beliefs. It found that these manipulations only had a small effect size. What’s more, it showed that the strongest effects happened directly after the manipulation instead of at the end of the experiment. This indicates that manipulation effects may not only fade after the experiment, but wane over the course of the experimental session – we don’t think about whether we have free will constantly. Not only that, but the review found that changes in free will beliefs did not have any meaningful, lasting change in behaviours.
A final issue has to do with replication, the confirmation of scientific findings in other studies. Whether a finding is replicated is one of the major principles in science. Repeating an experiment and consistently finding the same thing can increase confidence about an initial finding. And a replication crisis has recently engulfed psychology.
This is also the case for free will research. For example, some researchers have failed to change participants’ belief in free will. Others, despite successfully changing free will beliefs, have found difficulty replicating the consequences. For example, one study could not replicate the findings regarding cheating on the mathematical task. These difficulties question the credibility of the original findings and pose serious challenges to the assumption that negative consequences come about from a disbelief in free will.
Can a disbelief be good?
We have seen different reasons why it may be premature for researchers to conclude that disbelieving in free will has negative consequences. But we may also question whether these negative outcomes are to be expected anyway, and whether adopting a sceptical position can actually have societal benefits.
One view is that beliefs about free will make no difference to how people think and behave. Think about meaning in life, something which research has supposedly showed to be weakened by a disbelief free will. Imagine that a person’s meaning is strongly attached to their career. This person harbours the ambition to become a lawyer and after years of dedication, finally achieves her dream. Would the belief that she lacked free will affect her meaning and how she perceived this success?
Not necessarily so. She was still the one that got up early, arrived home late and dedicated hours to studying. The pride in her accomplishment does not evaporate in the knowledge that she was not the ultimate controller. She may be thankful for her upbringing, the education she received and her family support. There can still be meaning and gratitude. People ascribe meaning to all sort so things – intelligence, athleticism and beauty – none of which have got anything to do with free will.
What’s more, I conducted a study that specifically looked at meaning in life across people holding a range of views about free will. One group read a text passage denouncing free will (a traditional free will disbelief manipulation); and a second read neutral text (a control condition). I also had a third group which did not receive a manipulation but was selected by identifying as free-will sceptics.
All participants reported on how meaningful their life was. As would be expected, those manipulated to disbelieve in free will showed lower meaning than the control, supporting previous studies. But crucially, the sceptic group had the same sense of meaning and purpose as the control.
This indicates that, as suggested earlier, the manipulations claiming to make people doubt free will are not doing so in accordance with a sceptical perspective. More importantly, this suggests that the meaning in life of people taking sceptic views is unaffected by their philosophical position.
Further supporting this, a 2024 paper examined the views of free will sceptics exclusively and found that 41% (out of 164) claimed that their disbelief had no effect on their life at all. And the most reported consequences were indeed positive, including having compassion for others, being less controlling and more relaxed.
Does this mean a sceptical position could in fact provide societal benefits? One of the most debated consequences of free will beliefs concerns morality and responsibility. In the current legal system, retributive punishment is predicated on the assumption that the criminal was morally responsible for their action and acted out of their own free will. Supporting this view, research has found that participants with a stronger belief in free will are actually less empathic towards wrongdoers, and hand out harsher sentences. Yet, if free will does not exist, it would seem unduly harsh to punish an individual based on an act they were not responsible for.
Gregg Caruso and Derk Pereboom are prominent free will sceptics who question the morality of punitive decisions. They suggest that instead of punishing people because they are morally responsible for their actions, more empathetic, non-punitive approaches can be adopted in line with sceptical beliefs. For example, their quarantine model seeks to isolate offenders and provide rehabilitation until they no longer pose a risk to society – regardless of their level of responsibility. Seen in this way, a disbelief in free will could potentially overhaul a legal system that’s at odds with the view that free will does not exist.
Holding the belief that people can freely determine their circumstances may also lead to other harmful outcomes, such as social inequalities, injustice, a lack of empathy, hatred and victim blaming. For example, a belief in free will can promote the idea that poor people are somehow deserving of their abhorrent conditions and are living in situations of their own making. These perceptions can change if the notion of free will is relinquished. Because people are not morally responsible for their own situations, a greater deal of compassion and empathy could be shown towards those underprivileged. So there could indeed be positive outcomes of declining to believe in free will.
Researchers have claimed that making people disbelieve in free will can have dramatic negative consequences. A disbelief in free will may indeed lead to bad things; however, due to various methodological issues, it is premature to make this conclusion. In fact, it could turn out that disbelieving in free will is actually a good thing.
To hear about new Insights articles, join the hundreds of thousands of people who value The Conversation’s evidence-based news. Subscribe to our newsletter.
Tom St Quinton does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
A chimpanzee sits in the Pongoland enclosure at Leipzig Zoo, where a huge amount of research has been conducted since 2004.Marco Warm/Shutterstock
Leipzig Zoo in central Germany is a world-leading centre of great ape research. Recent studies have seen chimpanzees there using touchscreen controls to navigate virtual forests and locate food rewards – applying similar techniques to what they would use in the wild.
Other research (of which I was part) has investigated chimpanzees’ social curiosity. We discovered they actively seek out information about others’ interactions, even if it means forgoing food rewards. Keeping track of their peers’ latest social developments appears central to these great apes’ social wellbeing.
But in my decade working with Leipzig Zoo’s chimpanzees and bonobos, one question came up repeatedly. Were differences in how each great ape would cooperate and resolve conflicts simply down to its mood on a particular day? Or were there longer-term explanations – deep-rooted personality traits, for example, or their relationship history with other apes?
Long-term questions like this are very difficult to tackle in single studies, which often draw on just a handful of participants. So, my colleagues and I have developed EVApeCognition: a standardised database of 18 years’ worth of great ape experiences, decisions and relationships.
This lays the groundwork for answering many more questions about these extraordinary creatures’ cognition, intelligence and social behaviour. If a bonobo showed striking generosity towards a partner in 2008, for example, we can piece together whether that behaviour was linked with their stable disposition, a particular relationship, or some other factor.
Changing how we study great apes
In all, EVApeCognition comprises 262 experimental datasets from 150 scientific publications between 2004 and 2021. These were all overseen by the Wolfgang Köhler Primate Research Centre, headquartered at Leipzig Zoo. Eighty-one great apes participated in these studies, with the vast majority (78) taking part in more than one.
In the wild, great apes are social animals who live in stable groups with histories, hierarchies and relationships that change over time. In contrast, a large majority of the studies in our database were with apes in pairs that imposed strict control conditions.
So, moving to larger-groupstudies could offer a more ecologically relevant window on their cognition and social behaviour. Group settings can present apes with different problems that map more closely on to the social challenges they face every day in the wild.
Our most recent study, led by Kirsten Sutherland at the Max Planck Institute for Evolutionary Anthropology, found that great ape quartets maintained access to a pool of yoghurt for significantly longer than pairs did. Social tolerance played a key role, with more tolerant quartets maintaining access to the yoghurt for longer periods.
We found that cooperation was strongest when the highest-ranking individual showed restraint, emphasising the importance of tolerant leadership.
The new database also highlights an imbalance running through captive great ape research: chimpanzees dominate the record, while bonobos, gorillas and orangutans remain comparatively underrepresented.
Experimental performance does not occur in a social vacuum. A great ape’s willingness to cooperate on a task on a given Tuesday may not only reflect its intelligence, but whether it groomed its partner that morning, or if its status had changed within the group.
Providing this context is essential to understanding how everyday experience and social relations shape their cognitive development. Fortunately, the field is moving in promising directions, with the EVApeCognition database one piece of a larger picture.
The global ManyPrimates project, established in 2017, has already produced the most comprehensive overview of primate short-term memory. This shows that genetic lineage has played a larger role than ecology or sociality in the evolution of their short-term memory.
At the level of higher-order reasoning, we now know that chimpanzees update their beliefs by considering all sources of information before making a choice. A 2025 study showed they remained committed to an initial belief when counter-evidence was weaker, but revised this when the supporting evidence became stronger – a pattern long thought to be distinctly human.
Perhaps most ambitiously, the divide between captive and wild settings is also beginning to close. Research led by Sofie Forss at the University of Zurich, for example, has found a systematic “captivity effect” when presenting the same new stimuli to both wild and captive orangutans. The wild individuals responded far more cautiously to novelty than their zoo-housed counterparts.
Taken together, these efforts point in a common direction: toward an understanding of great ape cognition that is at once broader in scope, richer in context, and more faithful to the complexity of their social lives.
Alejandro Sánchez-Amaro does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
Login
.jpeg){kind=link}