Going Live, Staying Hidden: How Streamers Exploit the Information Gap

 

Livestreaming platforms such as Twitch, YouTube Live and Kick have grown to where creators broadcast in real time to millions of viewers worldwide. However, beneath the entertainment lies a fundamental economic problem: asymmetric information, where there is an imbalance of information across participants in an economic transaction. In this blog, we explore three branches of this issue - principal-agent problem, moral hazard and adverse selection - within the livestreaming industry.

 

In 2024, a live stream by Johnny Somali sparked fierce criticism across South Korea. While visiting the country, he kissed a statue of a "comfort woman," widely perceived as an insult to the victims of Japan's wartime atrocities (BBC, 2026). While many have condemned this act as outrageous, an uncomfortable question has emerged: why do creators engage in such actions?

 

To answer this question, you must understand one economic concept: the principal-agent problem. Platforms like YouTube and Twitch act as the principal, while content creators serve as the agent. Ideally, both parties can benefit from working together: creators produce content, and platforms provide exposure and revenue.

 

However, their incentives often conflict. Creators are rewarded for generating short-term attention. Platforms, on the other hand, prioritise long-term goals such as user trust, brand reputation, and content that appeals to advertisers. This conflict of interest creates tension. While harmful content can damage a platform's image, it can also generate high engagement in the short term. Feelings of anger lead to increased view counts, higher engagement, and ultimately, greater revenue. Therefore, his immoral strategies could be seen as economically rational from a content creator's perspective.

 

Furthermore, creators are not the only ones bearing the full cost of such behaviour. The repercussions, including public backlash, cultural harm, and damage to reputation, often fall on the community and the platform. Because creators do not face commensurate consequences, they take greater risks. As long as revenue is tied to view counts and engagement, the incentive to push the boundaries will not disappear. Why don't platforms simply put a stop to such behaviour? Whilst it is possible to impose bans or suspensions, doing so strictly risks reducing content supply and user engagement.

 

Ultimately, this highlights a fundamental contradiction in the digital economy: while platforms rely on creators for user engagement, they cannot fully control their behaviour. As long as attention remains the primary currency and the incentive structure stays unchanged, controversial content will continue to proliferate. This stems from a broader issue of asymmetric information, moral hazard, a situation that arises when one party in an economic transaction cannot observe another party's behaviour, leading to one party taking on excessive risk (Goolsbee, 2024).

 

The mechanism for this starts with revenue, which is a product of views, donations, and subscribers. In this current media climate, the most effective way to stand out in streaming is to do things outside the norm: extreme stunts, large-scale activities, or controversial content driving viewership. However, some streamers take this to the extreme.

 

Take Twitch streamer Kai Cenat, one of the platform's biggest streamers, with 19.8M followers and a peak concurrent viewership of 1M. In 2023, he did a PS5 giveaway in New York, livestreaming the event (BBC, 2023). Due to a lack of planning and coordination with the city, thousands of people rampaged through Union Square, hurling bottles, stones, and tins of paint. This caused the NYPD to declare a level four mobilisation with one thousand officers dispatched, assaulting police and resulting in sixty arrests. This suggests that Kai Cenat’s marginal private benefit was high through global viewership of this event, but his marginal private cost was moderately low, such as soft legal consequences, despite inciting a riot. The main costs were imposed on the public through property damage and injuries, the police, emergency services, and the platform. In microeconomics, this is called negative externalities, from the production of the live-streamer. As streamers do not internalise these external costs, risky behaviour exceeds the socially optimal level where marginal social cost is greater than marginal private cost.

 

After discussing the streamers, let’s turn our attention to the products they sell during their livestreams.

 

Have you ever wondered whether the products promoted by liverstreamers are genuine? They appear convincing on screen, and the discounts seem too good to resist.

 

In fact, such doubts are justified. A report by Reuters states that live-stream shopping makes it more difficult to identify counterfeit or substandard goods, as real-time video and rapid transactions limit effective verification (Reuters, 2024). This can also be attributed to asymmetric information. In the livestreaming market, sellers and streamers often know more about the products themselves than consumers do.

Consequently, consumers rely more on signals such as the streamer’s reputation to judge product quality. However, this often does not reflect the true quality of the goods. Whilst some streamers might promote high-quality products to safeguard their reputation, most streamers are typically incentivised to promote best-selling products to maximise commissions, rather than to recommend genuinely high-quality products. As a result, low-quality products can exploit the streamer’s reputation to be presented as equally trustworthy.

 

From the consumer’s perspective, products begin to appear similar in quality. When consumers are unable to distinguish product qualities, they might assign a price based on their perceived average quality. This might result in an unfair price for high-quality sellers, leading these sellers to exit the market. Thus, the average quality of remaining products declines, consumers become willing to pay even lower prices, and more high-quality sellers are driven away. This is known as adverse selection, a situation where there are stronger incentives for “bad” types of a product to be involved in a transaction than “good” types of the product (Goolsbee, 2024).

 

Overall, the livestream platform can highlight how asymmetric information can distort incentives across a wide range of areas, from creators’ behaviour to product markets. Unless these information gaps and misalignments of incentives are addressed, the market will remain inefficient, with society bearing costs that benefit only individual streamers.

References

 

Ng, K. (2026). ‘Johnny Somali: South Korea jails US YouTuber for public nuisance’, BBC News, [online] 15 Apr, Available at: https://www.bbc.co.uk/news/articles/cj400lje9vxo (Accessed: 23 Apr 2026).

 

Kim, C. (2023). ‘Kai Cenat: Twitch streamer says he's "beyond disappointed" after New York riot’, BBC News, [online] 4 August, Available at: https://www.bbc.co.uk/news/world-us-canada-66466699 (Accessed: 3 April 2026).

 

Goolsbee, A., Levitt, S. and Syverson, C. (2024). Microeconomics. 4th edn. New York: Macmillan Learning. Kortext. Available at: https://read.kortext.com/reader/pdf/948705/ (Accessed: 25 Feb 2026).

 

McLymore, A. (2024). ‘Focus: ‘Livestream’ shopping thwarts some high-tech tools to stop fake merch’, Reuters, [online] 8 May, Available at: https://www.reuters.com/technology/livestream-shopping-thwarts-some-high-tech-tools-stop-fake-merch-2024-05-07/ (Accessed: 24 April 2026).

The Billion-Dollar Lag: Why Ozempic's Successor Is Stuck in the Lab. A Microeconomic Look at Externalities, Incentives and Market Failure

 

Chances are, you've scrolled past at least one dramatic celebrity weight-loss story on social media recently. From stars like Oprah to Megan Trainer, everyone's favorite celebrity has seemingly attained their dream bodies overnight. These transformations didn't happen in the gym, but through the wonder drug: Ozempic. Ozempic, the weight-loss drug that has taken the world by storm, is the most well known product in a broader class of drugs known as GLP-1s.

Source: Daily Mail, 2024

 

GLP-1s work by making you fuller longer and curbing your appetite– simple in theory, transformative in practice. For patients, this means weight loss and significantly lower risk of heart attacks and strokes. But the benefits don't stop at the individual. A healthier population means less pressure on the NHS, which currently spends a staggering £98 billion a year dealing with obesity-related illness and lost productivity (Campbell, 2023). Employers benefit too as fewer sick days means more productive workers.

 

So far, so miraculous. But here's the uncomfortable truth: the very company behind Ozempic has no incentive to make it better, cheaper or more accessible. Novo Nordisk, the Danish pharmaceutical giant behind Ozempic, is sitting on one of the most profitable drugs in history. With demand already outstripping supply, why pour billions into researching the next generation of GLP-1s?

 

From a business perspective, the answer is simple: don't bother. The problem isn't greed, it's logic. The benefits of next-generation GLP-1 research, what economists call positive externalities, would flow to patients, the NHS and taxpayers, but Novo Nordisk would foot the entire bill. That's not a gamble any boardroom would take. And so the research stalls, and society pays the price. This is the “billion-dollar lag”.

When Profit and Progress Don't Mix

Source: Indiana University, 2023

 

Imagine you are on the board of directors of Novo Nordisk. Your current drug is a global phenomenon, demand is outstripping supply and the profits are rolling in. Now someone proposes funding the next generation of GLP-1 research, do you say yes?

 

Probably not. The odds of a new drug making it to the market are just ~15% (Schuhmacher, 2025). Furthermore, your patents could expire before the drug is even approved, meaning competitors can copy and sell it freely, leaving you with but a small portion of the returns on your own research. On top of that, even if your drug succeeds, much of the benefit flows to the patient, NHS and taxpayers rather than Novo Nordisk’s own balance sheet. In a gamble like this, you'd pocket the difference too.

This is what economists call the principal-agent problem. Society needs firms to take on research but cannot offer them enough of the reward to make it worthwhile. The result is a level of innovation far below what society actually needs– a market failure, plain and simple.

 

Why the Obvious Fix Isn't So Obvious

So what should governments do? The fix seems obvious. If companies won't take the research risk voluntarily, governments should make it worth their while. The two main tools governments can use to accomplish this are patents, which grant companies a temporary monopoly on a new drug, and direct research subsidies which offset the cost of research upfront.

 

Both options sound reasonable on paper, but they are not free. Patents are paid for by society through higher drug prices during the protected period, while subsidies draw directly from taxpayer funding. This means getting the balance right is not just a technical challenge but a question of how much society is willing to pay for innovation it may never fully control.

 

In a perfect world, governments could calibrate this precisely. Patents would last just long enough to reward innovation without blocking out competition. Subsidies too would cover exactly the risk needed to make research worthwhile- no more, no less. Companies would invest at the pace society needs and everyone would be better off. If only it were that simple…

Source: Authors' own

 

The reality is that pharmaceutical companies know far more about their actual research costs than any government does, that information gap, also known as information asymmetry, is where things unravel. Armed with that knowledge advantage, firms can lobby for subsidies well beyond what is genuinely needed, or push to extend patents far longer than is socially justified. This is why recent court battles over patent expiry such as Novo Nordisk's current case in Brazil, China and Canada are not one-off occurrences (IP fray 2026, Health Tech World 2025). Rather, they are what happens when the referee writes the rulebook.

 

Who Will Push Innovation Forward?

So where does this leave us? The “billion-dollar lag” is not simply a story of slow science or missing technology. It is a story of misaligned incentives. When existing drugs already generate billions in profit, further research looks like an expensive gamble rather than an urgent priority. And because the benefits of that research spill over to patients, the NHS and taxpayers rather than back to the firm, the market will never fix this on its own.

 

But it doesn't have to stay this way. When governments design intervention around the incentive problem rather than simply throwing money at it, the results can be remarkable. Operation Warp Speed, the US government's COVID-19 vaccine programme, promised to buy the vaccines before they even existed, removing the commercial risk that normally paralyses pharmaceutical drug research. The result was multiple approved vaccines in under a year (Gollom, 2020).

 

Applied to GLP-1 research, the same model could promise a market for next-generation drugs before they even exist, giving firms like Novo Nordisk a reason to take the gamble. The blueprint exists. The only question is whether governments will act before the next crisis forces their hand.

Reference List:

Campbell, D. (2023). Cost of people being overweight in UK now £98bn, study finds. The Guardian. [online] 4 Dec. Available at: https://www.theguardian.com/society/2023/dec/04/cost-of-people-being-overweight-in-uk-now-98bn-study-finds (Accessed: 23 April 2026).

Daily Mail (2024) 'Oprah Winfrey admits she DID use weight loss medication', Daily Mail, 13 January. Available at: https://www.dailymail.com/tvshowbiz/article-12860333/Oprah-Winfrey-admits-DID-use-weight-loss-medication.html (Accessed: 23 April 2026).

Gollom, M. (2020). Why Trump’s Operation Warp Speed is credited with helping race for COVID-19 vaccine. [online] CBC. Available at: https://www.cbc.ca/news/health/operation-warp-speed-trump-pfizer-moderna-vaccine-1.5806820 (Accessed: 23 April 2026).

Health Tech World (2025) ‘A key Ozempic patent lapses in Canada: Deliberate or a mistake?’, Health Tech World, 22 September. Available at: https://www.htworld.co.uk/news/a-key-ozepmic-patent-lapses-in-canada-deliberate-or-a-mistake-rg25/ (Accessed: 23 April 2026).

Indiana University (2023) 'Behind the Ozempic headlines', My IU, Available at: https://www.myiu.org/stories/behind-the-ozempic-headlines.html (Accessed: 23 April 2026).

IP fray (2026) ‘China’s top court grants Novo Nordisk win over key Ozempic patent’, ip fray, 5 January. Available at: https://ipfray.com/chinas-top-court-grants-novo-nordisk-win-over-key-ozempic-patent/ (Accessed: 23 April 2026).

Schuhmacher, A., Hinder, M., Brief, E., Gassmann, O. and Hartl, D. (2025) ‘Benchmarking R&D success rates of leading pharmaceutical companies: an empirical analysis of FDA approvals (2006–2022)’, Drug Discovery Today, 30(2), 104291. Available at: https://doi.org/10.1016/j.drudis.2025.104291 (Accessed: 23 April 2026).

Best Believe, British Beelines Bamboozles Boarders

 DingDong! Your train has been cancelled…

Have you ever found yourself at the station waiting for a train that never ended up coming? You could be waiting for one right now. But, you’re not alone! In 2025, Great Britain had 7.4 million trains planned and a 4.1% annual train cancellations score. This implies approximately 300,000 train cancellations left passengers stranded (Office of Rail and Road, 2026). We hear stories of the futuristic rail systems in Japan and Switzerland, where trains are affordable, on time, and rarely ever cancelled. Private operators, such as JR East, which have not raised fares in over 31 years, and the publicly owned Swiss railway system, where 60% of the funding comes from government subsidies and public investment (Harding, 2019). Yet, in the 6th largest economy, we pay absurd fares whilst having a 4.1% chance of never actually reaching our destination (Office of Rail and Road, 2026). Is this simply poor infrastructure and management? Or are private actors deliberately taking advantage of the principal-agent problem?

Figure 1: Mass Cancelling of trains (Getty Images, n.d.)

 

What is the Principal-Agent problem?

Train cancellations are not always a reflection of circumstances such as bad weather or strikes, they could be a principal-agent problem. The principal-agent problem occurs when an entity (the “agent”) is hired by and makes decisions on behalf of another entity (the “principal”). Since the principal can only observe the outcome, the agent will pursue their own interests rather than the principal’s. Here, the principal-agent problem leads to a moral hazard (Grossman and Hart, 1983). What’s a moral hazard, you say? I’ll leave that to future me to explain!

Figure 2: The principal-agent problem (Source: ChatGPT, 2026)

 

 

In the UK rail system, the government, as the principal, observes outcomes such as cancellations and delays. But, it cannot know how much effort train operators, as the agents, put into staffing, maintenance, disruption management, etc. In practice, train operators aren’t willing to reduce cancellations and delays by increasing spending on train operations. Put simply, if a company feels a decision is not worth the cost, it may decide to forego it. To try and solve this issue, the UK government has stopped renewing expired railway contracts and is moving towards nationalising the train system through establishing the Great British Railways (GBR) (Department for Transport and DfT Operator Limited, 2026).

 

 

Wait…but what do you mean by “Moral Hazard?”

Do you remember when I mentioned moral hazard? I’ll explain it now! Imagine you’ve hired a cleaner to tidy your home, paying them by the hour. When you’re home, they always appear to be hard at work, leaving every corner spotless. But as soon as you leave, they start slacking off, only mopping the floor just before you return.

Economists refer to this phenomenon as “moral hazard”. Once an agreement is signed, the agent can simply neglect responsibility since they don’t bear the full consequences. The agent may cut corners in their work, harming the principal.

 

 

The ones who made too many promises…

 

The UK rail franchising system has been in place since the 1990s, outsourcing passenger rail operations to private firms to introduce competition and improve efficiency. It sounds wonderful, doesn’t it?

 

 

Wrong! This system has created distorted, conflicting incentives. Operators know their own costs far better than the government, so to secure a franchise, they make extremely optimistic revenue forecasts, knowing that the government couldn’t verify this. A great example is Virgin East Coast, which promised to pay £3.3 billion to operate the East Coast line until 2023 (GOV UK, 2014), claiming that, in addition to more seats, passengers would enjoy cleaner, more reliable trains and faster journeys. However, within just a few years, the East Coast franchise failed, forcing the government to take over, using taxpayers’ money to foot the bill for a private company’s rash decision-making (Parliament UK, 2018). The franchising system created moral hazard twice. Operators inflated bids knowing the government would absorb the losses if they failed, then underinvested once contracts were secured knowing penalties were too small to hurt. The rail company knew that the government could not let the rail network grind to a halt if things went wrong. They enjoy the benefits of neglecting responsibility while others bear the cost. Who suffered when the franchise failed? That’s right: passengers like you left waiting for the train. It doesn’t get more morally hazardous than that! (Preston and Bickel, 2020)

Figure 3: Decision Tree Game

 

GBR: A new hope or the same old problem?

Next time you're stranded on a platform, remember that the system wasn't broken by accident. Private operators promised everything to win franchises, let the service quality decline through underinvestment, and then made taxpayers foot the bill when things went wrong. However, as the decision tree game above tells us, this isn't malice; it's rational economics. Play the game to find out how an agent can privatise gains and pass the losses onto you, the taxpayer. The GBR proposal to nationalise track and train could improve the situation, but the principal-agent problem isn’t resolved by just nationalising. Managers in a public rail body may still know more than ministers or passengers about costs, staffing, and maintenance, meaning weak monitoring can still lead to underperformance. GBR becomes the agent, meaning Parliament and the public become the principal. They must now hold a public body to account rather than a private one. Without binding and transparent performance targets, the incentive to cut corners doesn't disappear. Basically: new agent, same problem!

“Oh no! Should I give up?” - No! Here’s what YOU can do:

 

However, all hope isn’t lost. Switzerland’s SBB demonstrates that this problem is solvable. The Swiss Federal Railways' 1999 reform introduced what Desmaris (2014) calls a "net cost" model. Targets and budgets were fixed in advance through binding contracts with strict reporting requirements, meaning any overspend came out of SBB's own costs. The incentive to cut corners disappears because the state won't bail you out for promises you didn't keep, the agent bears the cost of their neglect. This is the model GBR needs. Write to your MP while the Railways Bill is still in Parliament. Contact the Office of Rail and Road. Ask what GBR's performance targets will be, and how they will be enforced. Demand accountability. Demand better standards. Demand a better GBR.

 

I hope our article was useful and informative! Why not take this short-quiz to see how well you understood it?

References

Department for Transport and DfT Operator Limited (2026) Great British Railways and the public ownership programme. Available at: https://www.gov.uk/guidance/great-british-railways (Accessed: 18 March 2026).

Desmaris, C. (2014) 'The reform of passenger rail in Switzerland: More performance without competition', Research in Transportation Economics, 48, pp. 290–297.

Getty Images (n.d.) Train departure board showing cancelled services in UK station. Available at: https://www.gettyimages.com (Accessed: 20 April 2026).

GOV.UK (2014) New East Coast mainline franchise confirmed. Available at: https://www.gov.uk/government/news/new-east-coast-mainline-franchise-confirmed (Accessed: 18 March 2026).

Grossman, S.J. and Hart, O.D. (1983). An Analysis of the Principal-Agent Problem. Econometrica, 51(1), p.7. doi:https://doi.org/10.2307/1912246.

Harding, R. (2019) 'Rail privatisation: the UK looks for secrets of Japan's success', Financial Times. Available at: https://www.ft.com/content/9f7f044e-1f16-11e9-b2f7-97e4dbd3580d (Accessed: 18 March 2026).

Office of Rail and Road (2026) Passenger rail performance. Available at: https://dataportal.orr.gov.uk/statistics/performance/passenger-rail-performance/ (Accessed: 18 March 2026).

OpenAI. (2026). ChatGPT Images 2.0 [AI image generator]. Available at: https://openai.com/index/introducing-chatgpt-images-2-0/ (Accessed: 23 April 2026).

Parliament.UK (2018) Rail franchising, House of Commons Committee of Public Accounts, Seventeenth Report of Session 2017–19. Available at: https://publications.parliament.uk/pa/cm201719/cmselect/cmpubacc/689/68907.htm (Accessed: 18 March 2026).

Preston, J. and Bickel, C. (2020) 'And the beat goes on. The continued trials and tribulations of passenger rail franchising in Great Britain', Research in Transportation Economics, 83, p.100846.

Fast Food, Slow Deliveries: Uncovering the Hidden Economics of Uber Eats

 

Seriously, why is my food taking so long?

We’ve all be there, stomach growling, staring at the Uber Eats map as our driver tales a suspicious detour or handles multiple orders. It’s easy to call it “poor service”, but it reflects a deeper economic issue behind the platform’s convenience.

The frustration comes from asymmetric information, as customers cannot observe how much effort drivers put into delivering orders. So, outcomes do not always match expectation, leading to delays and inconsistent service quality.

In this post, we explore how these hidden information problems lead to inefficiencies and why simple feedback tools don’t fully solve them.

Figure 1: Image of the order from Uber Eats (Group Member, 2026).

What’s actually going on behind your deliveries?

Have you ever tracked your food delivery on Uber Eats, only to notice your driver took a longer route or stopped multiple times along the way? While this may only seem like poor service on the surface level, there is a deeper economic problem concealed behind the convenience of such food delivery apps.

At the heart of the issue is a situation where one party in a transaction knows more information than the other, which is known as asymmetric information in economic theory. In this context, the driver knows exactly how much effort they are putting into delivering food, but both the platform (Uber Eats) and customers do not. Therefore, from the outside, the driver’s actions and efforts are not directly visible, suggesting that you pay for the service without fully understanding what is happening behind closed doors.

This produces what economists call a principal-agent problem. In other words, the platform Uber Eats, which is the principal, wants to provide fast and consistent food deliveries, while the driver, which is the agent, may have different objectives. Given that high effort requires time and energy, drivers may choose to avoid putting in high efforts to benefit themselves.

Therefore, once the driver accepts an order, their behaviour might change such that they may take longer routes or handle multiple orders at the same time, especially when their actions are not closely supervised which inevitably results in delayed food deliveries. This is known as moral hazard, where people change their behaviour when they realise they are not being closely monitored. This is evident from how “order staking” has become increasingly normalised in Uber Eats, often causing longer wait times for customers, and in some cases cold food which heaving damages customer satisfaction (Nott, 2024). This highlights the motivations of workers in gig economy platforms, such that they adjust their efforts depending on how they are rewarded (Skrzek Lubańska & Szban, 2024).

Overall, because drivers have more information about their own actions, they do not always fully bear the consequences of reduced effort, such as the reputation of Uber Eats amongst customers. This can lead to an inefficient outcome where not only delivery times are longer but also the quality of service is lower. To put it simply, the outcome is not as good as it could be when compared to the outcome if the efforts of drivers were fully observable. 

Figure 2: The GPS of the Uber Eat’s delivery (Group Member, 2026).

When Feedback Fails

A key way Uber Eats attempts to solve the problem behind your deliveries is by tipping and ratings acting as signals of driver’s effort and reliability, both would usually be unobservable. Without them there would be market inefficiency, where good drivers are not getting rewarded and poor service isn’t being punished.

Firstly, with a rating system, this allows customers to produce feedback of the drive to Uber Eats. This tries to improve allocative efficiency, as higher rated drivers are rewarded to “higher earning opportunities”, which incentivises drivers to work at a better quality (Uber, 2025). Additionally, the drivers are discouraged from low ratings as they “may lose access to the app”, which could act as a strong deterrent to drivers of Uber Eats (Uber, 2026a).

Rating systems may not be entirely accurate. Imagine a driver delivers a McDonald’s order on time and still warm, but McDonald’s forgets a milkshake they ordered. Even though the driver put their utmost effort into this order, the customer can give them a low rating for things out of their control. Uber Eats does have a policy in place to avoid misleading or fake reviews that are not based on “genuine experiences” (Uber, 2025b). However, there is still room for inaccuracies in the rating system, and so potentially Uber Eats needs an updated policy to better reflect true signals. This could potentially include allowing drivers to dispute low ratings with their side of the story.

Moving onto tipping, this can be considered a signal from customers of the overall quality of the driver. This brings back the idea that better service is rewarded, when both sides have equal information. Therefore, drivers who do go above adequately doing their job get financially rewarded, so this creates an incentive to do so.

In contrast, tipping varies from individuals, irrespective of effort (Saayman & Saayman, 2015). For example, one customer might tip largely for a McDonald’s delivery round the corner, but someone else didn’t tip anything despite the driver ensuring their fries weren’t cold. This means tipping doesn’t always reflect the driver’s effort and there is room for improvement with tipping as a signal for Uber Eat driver’s quality.

Overall, tipping and ratings can be considered as two signals for Uber Eats to lower asymmetric information. Despite this, due to the disadvantages of both signals, neither of them eliminates all of the unequal information about driver’s effort.

So, What’s Next?

In conclusion, although the efforts put in by the drivers are not fully observable in the context of food deliveries on Uber Eats, the solutions that were discussed in this blog have allowed these efforts to be incentivised to an extent. To further prevent poor service, such food platforms should further update their policies to better reflect a signal that would encapsulate the hidden effort of drivers.

References

Nott, G. (2024) ‘Cold food and poor pay: order stacking booming among food delivery apps’, The Grocer, 31 July. Available at: https://www.thegrocer.co.uk/news/cold-food-and-poor-pay-order-stacking-booming-among-food-delivery-apps/693908.article (Accessed: 12 April 2026).

 

Saayman, M. and Saayman, A. (2015) ‘Understanding tipping behaviour – an economic perspective’, Tourism Economics, 21(2), pp. 247–265. Available at: https://www.researchgate.net/publication/275528939_Understanding_Tipping_Behaviour_-_An_Economic_Perspective (Accessed: 6 April 2026).

 

Skrzek-Lubańska, M. and Szban, J.M. (2024) ‘Motivation in the Gig Economy: The Incentive Effect of Digital Platforms. A Literature Review’, Kwartalnik Nauk o Przedsiębiorstwie, 72(2), pp. 95–116. Available at: https://www.researchgate.net/publication/383969194_Motivation_in_the_Gig_Economy_The_Incentive_Effect_of_Digital_Platforms_A_Literature_Review (Accessed: 12 April 2026).

 

Uber (2025) Introducing the new Uber Eats Pro – better service, better status, better earnings. Available at: https://www.uber.com/us/en/blog/uber-eats-pro-preferred-deliveries/ (Accessed 6 April 2026).

 

Uber (2026a) How ratings work for delivery people. Available at: https://www.uber.com/gb/en/deliver/basics/tips-for-success/delivery-ratings-explained/ (Accessed: 26 March 2026).

 

Uber (2026b) Ratings Policy – Great Britain. Available at: https://www.uber.com/legal/en/document/?name=ratings-policy&country=great-britain&lang=en-gb&uclick_id=0bce070e-1317-49f1-b268-231963d2b630 (Accessed: 26 March 2026).

AI’s Hidden Costs: The Trade-Offs Behind the Tech

 Sacked by the Spreadsheet

Imagine waking up to find your job replaced by an algorithm. For millions of people this is now becoming reality; AI has redefined customer service and manufacturing in countless industries across the world (Sharp, 2024). While some marvel at the gains in productivity, there emerges a burning economic question: what are the hidden costs to progress?

The Price of Progress

In a free market, all participants in the economy freely exchange information as to their needs, wants, skills and abilities. This process creates an ever-updating network of prices that accurately reflect the relative costs and benefits of undergoing different economic activities. However, the efficacy of this mechanism often breaks down. Sometimes, prices fail to capture all the effects of producing a good or service. Economists call these hidden costs ‘negative externalities’ - when someone else ends up paying the price for a decision they weren’t part of (Kenton, 2024). If a firm replaces workers with AI, it may reduce its private costs by increasing productivity. However, the social costs of retraining workers and mental health crises caused by unemployment are externalised onto society. Thus, these costs are a classic case of a market failure, unaccounted for in corporate balance sheets.

Autonomous vehicles will destroy 500,000 jobs in California’s trucking industry, but Tech firms won’t pay the social cost (Garrifo, 2022). Displaced workers, increased demand for public support services, and the psychological toll on affected families are all effects that were not priced in by markets. Sound familiar? The process is much like factories polluting rivers without covering the bill for cleanup.

 

Convenience Over Competence

There are clearly social costs to AI that have been ignored - but are there also benefits that have been overestimated? Here’s where two classic economic ideas come in: moral hazard and the principal-agent problem. Moral hazard occurs when a party takes risks they ordinarily wouldn’t, because they don’t bear the full consequences of those risks. The principal-agent problem, on the other hand, occurs when one party (the agent) makes decisions on behalf of another (the principal), but their unaligned interests affect economic efficiency.

In many ways, AI is the time-saving tool people claim it is. There’s no doubt it stands out as a powerful method to generate refined signals that reduce information asymmetry in economic transactions. By embedding these signals into principal-agent contracts, moral hazard can be mitigated, boosting efficiency across various market structures. Take Uber as an example: AI tools help reduce asymmetric information by producing precise ‘effort signals’. For instance, cross-referencing a driver’s GPS data with assigned trips to see whether they are genuinely working or idling. If agents know their behaviour is being tracked closely, they have less leeway to shirk or misreport, reducing the moral hazard of not being monitored (Zhang, 2025).

The twist? AI doesn’t just fix moral hazard, it can create it too. As workers have increasing access to AI tools, they become increasingly reliant on them. Reliance on AI can affect workers’ own skills and pose potential problems if the AI software is not always readily available (Klingbeil, 2024). Let’s use an example that may cut close to the bone for fellow students. How many times have you asked ChatGPT for an explanation, copied it into your notes, and moved on?  This often means you’ve not really understood the topic. You miss figuring things out for yourself - making connections, thinking critically, or even just sitting with a tricky concept until it clicks. It might feel like you're learning, but you’re really just going through the motions.   

In economic terms, society doesn’t gain as much from AI as firms think they do, because the private benefits to firms outweigh the social benefit to everyone else. That’s where things get out of balance and may warrant action.

                                  

Policies and Protection

We’ve established that the unchecked adoption of AI is creating economic inefficiencies, because firms and individuals focus primarily on their private costs and benefits (Lane, 2021). Moral of the story? Sometimes the ‘invisible hand’ of the market needs a little nudge.

So how can governments and institutions intervene? Some possibilities include:

1.      Pigouvian Taxes: Financial penalties for firms fully replacing jobs with AI, negating the social costs of unemployment. This may realign the private and social cost of replacing workers with AI, ensuring a more socially optimal level of automation.

2.      Transparency Mandates: Require firms to disclose automation plans, empowering workers to adapt to AI developments. This could increase the productivity of workers, lowering the private cost of employing humans compared to automating the process.

3.      Mandatory Evaluation of Employee Skill Retention: Require firms to periodically evaluate potential erosion of employee skills due to overreliance on AI, mitigating overestimations of automation’s benefits.

Some governments have already acted. The EU’s AI Act (2024) set a precedent by classifying high-risk AI systems. Meanwhile, former US President Joe Biden’s AI Executive Order (White House, 2023) emphasises ‘worker-centric’ innovation. Nevertheless, more may be needed to ensure externalities are internalised - government action is not always enough. In theory, the market could sort all this out on its own, that’s what the Coase Theorem suggests. But in reality? Workers don’t have the power or information to negotiate with big tech firms, so it rarely works out that way.

Human-AI collaboration may offer a middle ground without intrusive or misguided government regulation. Germany’s 'AutoUnion 4.0' pays manufacturers to use AI to assist, not replace, workers. Without layoffs, BMW’s AI-assisted lines saw productivity rise by 25%. Evidence from California suggests autonomous vehicles might grow the economy while creating jobs and raising wages without mass layoffs (Wilkinson, 2023).

Balancing Benefits and Burdens

AI’s promise is genuine, but so are its distortions. When markets ignore hidden costs and inflate hidden benefits, progress has a serious price. It’s time to treat automation not just as a tool, but as a trade-off. It’s microeconomic tools like taxes and regulation that can help bring the balance society desperately needs.

References

1.      Brookings Institution (2023). Automation and Artificial Intelligence: How Machines Affect People and Places. Retrieved from https://www.brookings.edu/articles/automation-and-artificial-intelligence-how-machines-affect-people-and-places/

2.      California Legislature (2024). Senate Bill 1047 (Automation Transparency). Retrieved from https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202320240SB1047

3.      European Commission (2024). EU Artificial Intelligence Act. Retrieved from https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai

4.      Gariffo, M. (2022) Automated trucks could cost 500,000 US jobs, researchers say, ZDNET. Available at: https://www.zdnet.com/article/university-of-michigan-study-claims-automated-trucks-could-cost-500k-us-jobs/ (Accessed: 31 March 2025).

5.      Image 1, AI generated using ChatGPT 4o

6.       Image 2, Available at:(9) Five ways AI is transforming the trucking industry | LinkedIn

7.      Image 3. Available at:ChatGPT in education: How students and teachers can use AI to transform learning | YourStory

8.      Kenton, W. (2024) Externality: What it means in economics, with positive and negative examples, Investopedia. Available at: https://www.investopedia.com/terms/e/externality.asp

9.      Klingbeil, A., Grützner, C. and Schreck, P. (2024) ‘Trust and reliance on AI — an experimental study on the extent and costs of overreliance on ai’, Computers in Human Behavior, 160, p. 108352. doi:10.1016/j.chb.2024.108352.

10.   Lane, M. (2021) ‘The impact of artificial intelligence on the labour market’, OECD Social, Employment and Migration Working Papers [Preprint]. doi:10.1787/7c895724-en.

11.   Lane, M. and A. Saint-Martin (2021), “The impact of Artificial Intelligence on the labour market: What do we know so far?”, OECD Social, Employment and Migration Working Papers, No. 256, OECD Publishing, Paris, https://doi.org/10.1787/7c895724-en.

12.   Sharps, S. (2024) The impact of AI on the labour market, Tony Blair Institute for Global Change (TBI). Available at: https://institute.global/insights/economic-prosperity/the-impact-of-ai-on-the-labour-market (Accessed: 31 March 2025).

13.   White House (2023) What they are saying: President Biden issues executive order on safe, secure, and trustworthy artificial intelligence | The White House, National Archives and Records Administration. Available at: https://bidenwhitehouse.archives.gov/briefing-room/statements-releases/2023/10/31/what-they-are-saying-president-biden-issues-executive-order-on-safe-secure-and-trustworthy-artificial-intelligence/ (Accessed: 02 April 2025).

14.   Wilkinson, L. (2022) Autonomous long-haul trucking stands to grow the Golden State’s economy while creating jobs and raising wages without mass driver layoffs, Silicon Valley Leadership Group. Available at: https://www.svlg.org/study-shows-autonomous-trucking-will-grow-californias-economy/ (Accessed: 31 March 2025).

15.   World Economic Forum (2023). The Future of Jobs Report 2023. Retrieved from https://www.weforum.org/reports/the-future-of-jobs-report-2023/

16.   Zhang, T. and Zhang, Y. (2025) Generative AI and information asymmetry: Impacts on adverse selection and moral hazard. Available at: https://arxiv.org/html/2502.12969v1 (Accessed: 31 March 2025).