Social Sciences

On the 22nd of June 2026, UK prime minister Keir Starmer confirmed the much anticipated news that he will be resigning from his role as leader of the Labour Party and therefore prime minister of the United Kingdom, effective after a leadership contest which will decide who is appointed as the UK’s new prime minister. However, the process to now appoint a new prime minister after a resignation is different to such methods that occur at a general election. 

In a typical general election process that most of us will be familiar with, each political party, such as Labour, Conservatives, Reform, Liberal Democrats and the Green Party will choose a candidate to run in a constituency across England, Scotland, Wales and Northern Ireland. A constituency is a local area whose voters that live within the area elect one MP (member of parliament) to represent them in parliament. Whichever party wins the most constituencies, or seats, typically wins the election and has their respective leader become the UK’s prime minister. 

However, in cases such as the current situation in British politics as a result of Starmer’s resignation, a different process is followed. The prime minister officially resigns by formally informing the monarch at the time, currently King Charles III. Then, the party in power (in the current case Labour) begins a leadership election to choose a new leader for the party and therefore new prime minister. However, as this process takes time, the outgoing prime minister will remain in office to ensure that the government can continue to operate properly during this changeover period. 

MPs vote for their preferred candidate for their party leader in a different way to how the public vote for their preferred party. Rather than voting for one singular party once, MPs from the party vote in a series of rounds to narrow down the candidates. Typically when two candidates remain, the final vote is decided by the party in power’s members across the country. The winner of this deciding vote becomes the new party leader. Many predictions have suggested that recently elected MP for Makerfield, Andy Burnham, could potentially follow this path and become the new Labour party leader which would result in him being the newest prime minister. 

As previously mentioned, this takes a lot of time to be organised, carried out and finalised, which means that the outgoing prime minister, Keir Starmer, will remain in office until the new labour party leader is elected. Starmer had said in his resignation speech that he would like this process to be completed by the time parliament ‘returns’ in September, around two months away. Should this timeframe be adhered to, in September, the new Labour party leader will become the new prime minister, and Starmer will officially leave his role as prime minister. 

When the new party leader is decided upon, they are invited to the reigning monarch’s official residence, currently King Charles III at Buckingham Palace. The monarch, King Charles III, will ask the new party leader to form a government, as due to being recently elected by their respective party members, the new party leader will be considered the most likely person to command the confidence of the House of Commons. When the new leader accepts this invitation, they are officially appointed as the Prime Minister of the United Kingdom. 

However, the process is not yet completed at this stage. The newly appointed prime minister must form his/her cabinet. The cabinet is a group of the most senior government ministers who help the prime minister run the country and make important decisions. These cabinet members are officially appointed by the King, advised by the prime minister, in order to respect tradition. When the cabinet is selected the new government begins work and attempts to maintain the support of a majority of MPs in the House of Commons, in order to support a stable government.

This exact process will be taking place over the coming weeks and months after Starmer’s resignation, so keep an eye out for any of these processes being reported on in the news!

The resignation of Keir Starmer this morning marks yet another dramatic turning point in British politics. With the former Health Secretary, Wes Streeting, already signalling his intention to stand in a leadership contest, and Andy Burnham returning back to Parliament and widely expected to challenge, the Labour Party is preparing for a rapid and competitive leadership race. If a new Prime Minister is appointed in the coming weeks, it will be the seventh change in national leadership since 2016, highlighting a profound shift away from long‑term political stability and towards a pattern of constant turnover.

One of the most significant drivers of this instability is the rise of gaffe culture — a media environment in which minor mistakes are magnified into defining political moments. While individual gaffes rarely shift public opinion alone, the cumulative effect of relentless hyper‑scrutiny can be deeply damaging. Rishi Sunak’s decision to leave a D‑Day commemoration early is a clear example: although seemingly a small misjudgement, it dominated headlines for days and contributed to a sharp decline in his approval ratings. According to YouGov polling at the time, Sunak’s net favourability fell to –51%, one of the lowest scores recorded for a sitting Prime Minister. Similarly, Ed Miliband’s infamous bacon‑sandwich photograph became a symbol of perceived awkwardness, which splashed across front pages with headlines like “Save Our Bacon.” These moments illustrate how British political journalism often prioritises image over policy, creating a culture where optics overshadow competence.

This dynamic was evident throughout Keir Starmer’s premiership. Despite overseeing measurable improvements in key economic indicators, with the UK recording 1.1% GDP growth in the last quarter, inflation falling to 2.3%, and unemployment dropping below 4%, Starmer still struggled to connect with the public. His approval ratings consistently lagged behind his government’s performance. At one point, polling placed his personal favourability at –40%, lower than several deeply unpopular public figures. Media portrayals frequently emphasised his “wooden” delivery and perceived lack of charisma, shaping a narrative of a leader unable to inspire or emotionally engage the electorate. Some commentators argued that Labour’s long‑term success depended on replacing him, as the party’s achievements were being overshadowed by his personal unpopularity. His resignation today reflects the growing political reality that image, relatability and media performance can be as decisive as policy outcomes.

Another key factor behind Britain’s rapid leadership turnover is the power of media framing combined with the relentless 24‑hour news cycle. Once a dominant narrative takes hold — that a leader is “losing authority,” “out of touch,” or “on the brink” — it spreads rapidly across broadcast media, online platforms and social networks. Continuous coverage leaves leaders with almost no time to recover from mistakes or shift public perception. Political crises that once unfolded over months now escalate within hours. This accelerates internal party pressure, fuels speculation, and creates an environment where leaders are judged not on long‑term governance but on their ability to survive each news cycle. In this landscape, leadership becomes fragile, reactive and increasingly short‑lived — a defining feature of modern British politics.

Taken together, these trends reveal a political landscape increasingly shaped by media‑driven volatility rather than long‑term governance. The rapid turnover of Prime Ministers, culminating in Keir Starmer’s resignation this morning, reflects a system where leaders are judged less on measurable outcomes and more on how effectively they navigate a relentless, image‑focused media environment. Economic indicators, policy achievements and legislative progress struggle to compete with the speed and intensity of modern political coverage. As gaffe culture, hyper‑scrutiny and the 24‑hour news cycle continue to dominate public perception, Britain faces a fundamental question about the future of its democracy: whether meaningful leadership can endure in a climate where narratives shift by the hour and political survival depends as much on media performance as on policy. Confirming details with trusted news sources is always essential, especially during periods of rapid political change.

Introduction

Further exploring explanations for behaviour may not only alter our perception of criminal behaviour, but the decision making that sentences it. Not only does this implication of determinism influence the jury that determines verdicts, but also the legislators and judges that govern the system who have naturally biased thought processes and subjective human experiences. It is however well established that the legal system is retributive, relying on the moral responsibility of agents believed to have free will over their actions. Therefore, as the determinism explored by psychology and neuroscience becomes increasingly influential, criminal sentencing may not only become more rehabilitative, but every process within the legal system will be more understanding of the impact subconscious bias has on fair and objective outcomes.

Free Will Versus Determinism

A prevalent discussion within most disciplines that explore or depend upon understanding of human behaviour is the debate between free will and determinism. Within psychology, the foundational debate between nature and nurture usually precedes the ultimate conclusion that interactionist approaches, considering a plethora of factors, contain the most representative and accurate depictions of complex human behaviour. Therefore, accepting that biological influences, like genetic predispositions, and environmental influences, like adopting the principles of those around you, contribute to motivation and decision making processes. Taking the holistic approach of both free will providing moral agency along with the context of determinism predetermining motivation and opinions can be described as representing the philosophical theory of compatibilism. This allows for the betterment of society when we take responsibility for our actions and decisions, and can explain behaviour for social development and prevention. For example, currently prevalent implications of polarising social media algorithms and online environments that perpetuate misinformation or extreme viewpoints may be a growing cause of concern when acknowledging the social influence this has on decision making. A sign of a healthy democracy is a functional and just legal system; however, (for example in the most serious criminal cases) if the main determiners of guilty verdicts are predisposed to potentially extreme or biased opinion, democracy may be eroded. So, it is important to explore how public opinion can be predetermined and how the negative consequences can be mitigated to ensure democracy.

Bias Determining Verdicts

Preventing the influence of external factors on decision making requires an understanding of the sources of influence people are susceptible to, and may be subject to throughout the course of justice. Confirmation bias seems to be a major factor, where individuals are more likely to value evidence they agree with, for example, agreeing more with the prosecution if it confirms their previous convictions. Research suggests that we have a rapid, involuntary bias to positively view statements we already agree with (Jarrett, 2018). Crucially, the principle of “innocent until proven guilty” could be undermined as a result of legal bias as verdicts are not decided purely based on evidence. Other examples of cognitive biases include what is known as the Halo Effect. This is a cognitive bias that occurs when an initial positive perception of a person alters the overall perception of the individual in accordance with the initial impression. It becomes difficult to alter the initial impression made when new or opposing information is presented. This may also implicate the influence of initial perception, positive or negative, transferring to judgement of an individual's actions (Psychology Today, 2019). This suggests that cognitive biases can influence verdicts. For example, when a jury already agrees with one party, it is more likely to be difficult to change their minds when presented with alternative evidence.

Solutions

While there are current solutions in place, further bias could be reduced through necessary improvements. For example, it is presupposed within the law that every judge has the capacity to determine their own biases for disqualification. Legal bias may be introduced when judges draw on their personal experiences to make judgments with limited and subjective experiences, potentially preventing their capacity to remain open minded, making it difficult for judges to mitigate their own bias due to it being predominantly subconscious. 

The nature of human biology explored in behaviourist approaches to psychology highlights the inevitability of experiences containing conflicts of interest on informing verdicts, emphasising that the disqualification of judges to prevent extraneous factors and error is crucial to consider carefully. To resolve issues and inconsistencies, policy-making processes could be used that balance public opinion that judges could be disqualified for a broader range of cases, legal policy considerations and the science on human decision making - allowing for dialogue between the public, the legal community, and psychologists (Higgins and Levy, 2020). Furthermore, while there is the option to examine bias through background checks, there is also potential for innovative ways of reducing overall bias in the decision making. 

For example, a jury could be created where the the most extreme biases are removed but the rest contain a representative pool of biases. Potentially exemplified by intentionally incorporating a variety of individuals with and without faith in the justice system (Curley, Dror and Munro, 2022). Juror bias presents a threat to democracy because it is important for decisions to be made based on evidence rather than prejudiced opinion or misinformation, and this could be achieved through methods like a change in the sampling method to limit the influence of bias through representation. Overall, developing policy and methods of selection could reduce bias contained within varying components of the legal system. Techniques to improve future outcomes may also apply to attempted implementation of Artificial Intelligence into the legal system. Biases can occur within this in the processes of data collection and labeling, model training, and deployment, leading to many types of bias including selection, confirmation, and stereotyping bias (Chapman University, 2025). Based on suggested solutions to human bias, intentionally deriving information and analysis from sources perhaps potentially biased in different ways could reduce the influence of predisposition based on biased information and algorithms.

Conclusion

Overall, it is evident that as deterministic influences become more widely understood and accepted, the legal system has to adapt with this knowledge to ensure it is as fair and just as possible. Understanding current mitigation for bias can allow legal systems to prevent future harm before social influence can perpetuate it. Evidence based policy and jury selection could be useful tools for ensuring justice despite the inevitability of bias and socially influenced opinion. In a time of rapid change, it is more important than ever for the leaders of governing systems to adapt when presented with new challenges.

In modern society it is undeniable that one of the key factors driving our future is the development of AI (artificial intelligence). Its uses are already reshaping the way in which multiple major industries operate, including accelerating medical advancements, controlling scheduling and research in companies, along with algorithms being used to fine tune employment regimes and using the modelling of data to predict events.

Given the large positive influence that it potentially holds it makes sense that creators and users of the software are trying to push the boundaries of AI to harness its capabilities. This involves developing it to have independent thinking abilities which allow it to outperform and exceed the abilities of human counterparts in industries and maximize the contribution that it has to society as a whole. 

However due to this unrestricted freedom given to AI it leaves the possibility open that they can commit acts within society that can be considered crimes, such as fraud, which can in theory be committed independently by the programme without the input of the developers.

This has led to a debate within the legal community as to how the law should be structured in order to prosecute in these events. It is interesting because there doesn’t appear to be a clear cut solution to this problem as all opinions will rationally favour a certain side but produce an equally strong counter argument for the opposition. 

For example, you may simply think that whatever specific software model that committed the crime should be discontinued or heavily restricted as a result of its actions. Whilst you could say this prioritises the safety of society, it can be just as easily justified as an unfair judgement as we wouldn’t punish all humans for a crime one person commits, therefore, how would it be fair to punish a particular software when there are all the other AI models which have contributed nothing but positively to society. 

The primary issue here is that despite being from the same software it doesn’t necessarily mean all AI models from there will think the same, they likely will generate individual outputs as a result of the general dynamic thinking given to AI, much like whilst we are all humans we do not all think the same things. 

The current situation is that the developers of the AI are the people prosecuted for any crimes committed by the software, due to AI lacking mens rea, which is the mental state required to commit a crime with malicious intent. As a consequence of being judged to lack the capacity to intend malicious acts this results in human developers being held legally accountable for the crimes, whilst simultaneously creating this culpability gap.

The culpability gap can be summarised as the AI causing damage in society despite the human developers who are punished lacking any intention for this to happen, therefore being prosecuted in spite of it being unclear ethically to whether they are liable for the software’s actions. 

By looking at key factors assessed in criminal law cases we can better understand the difficulties produced when determining criminal liability. Liability is dependent on two primary components, actus reus, which is the guilty act and mens rea, which we have already briefly looked at, which is the guilty mind.

Actus reus must be a voluntary action which is prohibited resulting in a consequence, or conversely being in a prohibited situation without a voluntary act, for example being found in possession of an illegal item such as a gun.

Mens rea is the psychological component of the crime, which can be established by either having malicious intent, showing recklessness or displaying negligence. By showing one of these elements during a crime you can be considered morally culpable.

To be considered guilty for a crime, both the actus reus and mens rea must be proven beyond reasonable doubt and they must also be shown to have occurred at the same time. If the two are disjointed then it doesn’t provide sufficient evidence of guilt. 

This was seen in the case Fagan vs Metropolitan Police Commissioner, where Fagan unintentionally drove his car over the officers’ foot. When the officer addressed Fagan, he swore at the officer and refused to move his vehicle. Despite now having the intent to harm he wasn’t performing any new action as he had developed the mental intent to harm after the initial act.

Despite these technicalities they are usually managed contextually to the case by the courts, which was shown in this case as Fagan was still found guilty under the grounds of the crime being considered a continuing act, although technically the actus reus and mens rea occurred at different times.

Now that it is clear what is judged as liable in criminal procedures, it becomes more obvious why artificial intelligence isn’t prosecuted as it lacks both components of actus reus and mens rea, being involuntary to its actions, which are considered as outputs of coding and they lack the human consciousness to actively cause harm.

On the contrary this can be considered as a relatively short-sighted stance, as a result of the rapid advancement of AI it brings into question the threshold required for the two components, as to what extent can an independent mind be considered to have the mental capacity to commit a criminal action voluntarily. 

At the trajectory AI is evolving it would be naive to dismiss the possibility of these software developing a consciousness which would rival that of a human. Although this may never happen as the programmes are designed to provide intelligence which involves the processing of information rather than the development of actual emotion/feeling. 

However the conscience is not an exact science and humans still do not have a concrete answer to what the source of it is. This ultimately leaves a chance that AI can be programmed to such a sophisticated degree that it could potentially develop a similar conscience to humans and that would drastically change the legal landscape when approaching these cases.

Reverting back to the current legal situation, the structure of liability in these cases involving AI is very complex and case dependent. This is due to the various ways in which AI is used to commit crime, to which we have only currently considered when the software independently chooses to commit the crime without human intervention. 

At the moment this crime is considered theoretical due to AI not being recognised in the legal system as being liable for autonomous offences, therefore there is no figure for the number of crimes caused by it. These are currently determined as human negligence crimes and are extremely rare, although arguably the most concerning circumstance of AI crime.

Human misuse currently accommodates for almost all recognised cases of AI crime, where the software is used as a tool by humans to assist them in committing criminal offences. 

In cases where an individual user is responsible for the input that results in a crime, that user is considered liable for the action. This is known as perpetration by another. This is typically seen when AI is used to generate deepfakes to commit fraud.

If the model is naturally and foreseeably programmed to consequently break the law, then developers and programmers are held responsible for the crimes, which is branded as aiding and abetting. An example of this is when the software is provided to a third party to participate in cyber attacks.

Finally, organisational negligence in a business setting occurs when AI is misused, the focus will initially shift towards the company who will be evaluated as to whether they have broken codes of conduct for AI . If this is disproven then the developers of the software are once more considered accountable for its actions, as they can be penalised for developer negligence.

In order to follow the legal framework, companies must adhere to the dual concept of digital penal compliance.

Firstly, this involves recognising AI can commit crimes and the company will be held liable for its actions, therefore in cases where breaches in the law are seen the company will be forced to update its compliance programs in order to show they have addressed organisational fault.

Secondly, it involves the use of AI as a tool to achieve compliance, where it is used to detect crimes within the company to implement anti-crime initiatives, rather than solely being seen as a criminal risk that the company has to manage.

An example of the legal statutes organisations must follow is the EU AI act which penalises systems based on societal threat. This includes banning systems which are considered actively dangerous to livelihood and strictly regulating systems which are utilised in critical infrastructure, such as employment, but require detailed risk assessments and management. 

On the other hand many systems are not seen as a risk and are unregulated, seen through video games and other systems which involve human interaction like chatbots, viewed as having limited risk, although users must be aware they are conversing with AI.

The problem which is still an issue is the nature of how the law addresses AI crimes. This can be done through two main updates to the legal framework, which either focus on holding humans or software accountable. Likely there would need to be an integration of the two punishments as it wouldn’t be fair to punish one of the two objectively of the situation.

A solution to address human involvement would be to focus on expanding the law's civil liability rather than criminal law in these cases where no one is objectively at fault, which would mean the developer/operator would be forced to provide compensation to the victim, without being held criminally liable.

In more malicious cases where a human is evidently at fault, then the person involved could be held to strict liability to act as a deterrence for intentionally misusing AI.

To address AI, whilst still not having any conscience, it would be sensible to continue progressing regulatory acts, which would sanction dangerously coded software including deleting certain programmes or having them in highly limited usage only available to certified individuals.

In conclusion, AI should be used to advance society and we must make adaptations to allow it to be part of our long-term future. This naturally involves making amendments to current legislation which currently focuses only on punishing humans involved in crimes. This is not always easy to prove due to AI having independent thinking, making it difficult to display the human intent to commit the crime in some cases.

Enforcing strict liability would likely provide a stronger deterrence in these cases as there would be less leeway for humans to misuse AI and not be held accountable. However, the most interesting circumstance would be if AI commits crime without human input. This would likely be handled by expanding civil liability, which seems the most ethical decision, as no one is directly responsible and the victim would be appropriately compensated.

Whilst AI lacks the actus reus and mens rea to be considered criminally liable these changes would be the best course of action to help solve issues relating to AI crime. In theory if AI does ever develop a proven conscience, although unlikely, then more alterations in the law would be required in the approach to AI crime, as if that was the case they would have all the components to be considered criminally liable for their actions.

Ignoring that possibility, by making changes in present legislation, then many of the issues surrounding the punishment for AI crime would be resolved or greatly supplemented, by creating more transparent laws surrounding criminal liability's relationship with AI.