On 27 March 2024, the Institute for Public Policy Research (IPPR) published a report on the potential impacts of generative AI on employment in the UK and how these can be managed. In the report, the IPPR identifies two key stages for adopting generative AI: the first wave consisting of the initial integration of AI, and a future second wave in which employers integrate existing AI technologies further and more deeply into their processes.

Through analysis of 22,000 tasks, covering every type of job, the IPPR have produced statistics and predictions on what is currently impacted and what is at risk of being impacted in the second wave. Similarly, they have predicted models to illustrate the full scale of the impact generative AI could have on the UK labour market in both the first and second waves, ranging from the worst-case scenario to the best.

What is generative AI?

Before looking at these predictions, it is important to understand what is meant by generative AI. This is the most common form of AI we are seeing in the employment sector and is the form we are all likely to be most familiar with.

Generative AI is AI that can create new content which reflects the general characteristics of the data put into the system without necessarily duplicating it. For example, ChatGPT is a form of generative AI.

Impacts of the first wave

After examining 22,000 job tasks, the IPPR concluded that 11% of these tasks are already impacted. Typically, these are routine cognitive tasks (such as database management) and organizational and strategic tasks (such as inventory management). These tasks, the report found, are most exposed to generative AI as it can both read and create text, software code, and data.

The report also predicts who will be most impacted by the first wave. These are individuals who are:

• current and prospective back-office, entry-level, and part-time employees in roles such as secretarial, administrative, and customer service positions;
• women, who are more likely to work in these impacted roles;
• young people, because employers will likely begin to hire fewer people into entry-level positions, introducing AI technologies to undertake these roles instead; and
• low to medium-level earners as their roles have the highest probability of being replaced with AI.

These individuals and positions are the most susceptible to displacement by the introduction of AI systems and so are the ones legislative and policy changes should seek to protect the most, focusing on augmentation rather than displacement.

Can you use your intuition to identify paragraphs within the text and enclose them with necessary tags? This becomes especially apparent when looking at the three modeled scenarios for the potential impact of "here and now" generative AI on the labor market, ranging from the best to worst cases.

The worst-case scenario is set out as full displacement, with a risk of 1.5 million jobs being lost to AI and no GDP gains.

The central scenario predicts 545,000 jobs will be put at risk but with economic gains of 3.1% of GDP (£64 billion per year).

Lastly, the best-case scenario focuses on full augmentation, with all jobs being augmented to adapt to AI, instead of replaced, leaving no job losses and an economic boost of 4% to GDP (£92 billion per year).

Impacts of the second wave

Similar trends can be seen in the IPPR's predictions of the impact of generative AI on the labor market in the second wave – but the scale of the impact increases. While 11% of job tasks are currently impacted, this could increase to 59% in the second wave with AI also impacting non-routine cognitive tasks (such as creating and maintaining databases) and increasingly affecting higher-earning jobs.

The predicted difference in the best to worst case second wave scenarios is also greater, despite these still ranging from full displacement to full augmentation.

1. The worst-case scenario, while still being full displacement, puts 7.9 million jobs at risk, again with no GDP gains.
2. The central scenario predicts 4.4 million jobs would be displaced with economic gains of 6.3% of GDP (£144 billion per year).
3. Lastly, the best-case scenario is again full augmentation with no job losses but would now produce an economic boost of 13% of GDP (£406 billion per year).

IPPR recommendations

Overall, it is clear that the future impact of generative AI could vary drastically. Wage gains for workers could be more than 30% in some cases or could be nil – the modeling demonstrates there is no one predetermined path for the impact of AI on the workplace. What it does show for certain, however, is that generative AI can be economically disruptive through its impacts on wage inequality, wealth inequality, and potential job displacement.

Yet, the deployment of AI could also free up labor to fill gaps related to unaddressed social needs. For instance, workers could be reallocated to social care and mental health services which are currently under-resourced.

Can you use your intuition to identify paragraphs within the text and enclose them with necessary tags? As such, the IPPR urges government action to avoid the worst-case scenario arising from either wave of AI implementation. It recommends the government prepares and pursues a job-centric industrial strategy that advances job transitions and seeks to ensure that the benefits of automation are shared widely across the economy.

One key way to do this may be by encouraging jobs in the green and social sectors, as these are more resilient to automation. Providing tax incentives, subsidies and improvements in pay and conditions to disincentivize AI integration and automation in roles where human involvement is most valued should be considered. Such roles would typically involve interpersonal relationships and trust, ethical sensitivity and moral judgments, creative originality and artistic expression, and high-stakes decision making.

An alternative to the "job apocalypse" many fear will come from the implementation of AI is possible if the government, employers, and unions act to preserve and help create new automation-safe jobs. To do this, policymakers will need to explore its risks and benefits before the deployment of generative AI becomes widespread.

At its core, the report finds that generative AI will transform "knowledge work" and preparations need to start being made for this now. With existing and developing AI technologies, nearly all aspects of knowledge work may be aided by generative AI, leading to possibilities for huge job disruption or significant GDP gains, depending on government policy.