AI Impact 2025

We asked members to give their views on which sectors they felt AI would have the greatest impact in 2025. Responders could choose only one option.

This is what they chose, in order:

For the top three rated areas, we asked an LLM to provide an overview of the reasons members gave as to why the respective area would be most affected by AI in 2025. We have also drawn out some specific comments, in the commenter's own words, to give a flavour of the feelings behind the thoughts.

Information Technology

1. Skills shift: AI will continue to lead to job reductions for mid-level programmers. High-value, highly paid roles and low-level jobs will remain but be impacted by automation
2. AI as a tool: while AI is transformative, it is not yet fully intelligent and is currently seen more as a tool for improving services and tasks, especially in IT
3. Driving R&D investment: the adoption of AI will significantly drive research and development budgets, focusing on AI's potential across IT sectors
4. Software engineers and IT support automation: AI is expected to replace many roles in software engineering and IT support by automating tasks, streamlining processes, and reducing the need for manual intervention
5. Early adoption: IT companies are acting as guinea pigs for early adoption, exploring AI to enhance productivity and efficiency. As AI matures, more integration will happen across industries
6. Infrastructure and service provision: AI can help manage complex IT infrastructures, like multi-region on-prem and cloud setups, and contribute to better service provisioning through deeper AI integration
7. Improvement in development: AI aids in code generation, documentation, and software development, providing efficiencies and enabling quicker delivery times. Many foundational elements in IT are already AI-ready
8. AI threat detection and security: AI will play a crucial role in threat detection and cybersecurity, helping safeguard systems and obfuscate attacks
9. Positive ROI in IT: Companies already see a return on investment (ROI) with AI applications in IT, particularly in software development tools and productivity aids
10. Workforce adaptation: IT professionals, who have the necessary skills, will be better positioned to adopt AI tools effectively. AI will improve workforce productivity by automating repetitive tasks, thus lowering headcount needs.
11. Increased AI adoption: As AI matures, it will continue expanding across IT processes, from code writing and documentation to overall productivity enhancements in the IT industry
12. Infrastructure and data management: AI will automate data analysis, manage large data volumes, and assist in design and coding across IT systems, driving efficiency and innovation

Several commenters noted that as IT underpins AI, it will be the first, and most, affected. One IT professional comments: ‘While it may be that the current generation of LLMs and "traditional" ML models would be easier to implement and far more disruptive in other industries, IT and software development are unfortunately for them the only industries generally capable of understanding and implementing meaningful AI solutions. I expect other industries to remain undisrupted in the short term while the barrier for entry to "AI" drops or they improve their technical competency.’

The demands of the end-user can drive big changes, as one member notes: the ‘move to increased personalisation is a big agenda item. Consumers want to spend more of their own time on things that are important to them. Tech alone does not make a person function, however, tech and AI can be fabulous life-enablers.’

There are still reservations for some. As one member comments: ‘Solid business use cases remain infrequent. In the majority of cases, true AI is a "tool to play with" rather than having a solid set of requirements and a meaningful financial business case.’

And the ever-present human-in-the-loop concern came up, with some helpful caveats. One member says: ‘A lot of the grunt work can be done by AI in requirements and coding. People can add value by refining the results’, with another simply writing ‘AI is perfect for creating suitable interfaces into IT.’

Health and social care

1. Ageing population & workforce shortage: an ageing society is driving the need for new tools, as there aren't enough trained professionals to handle the growing demand. AI can help bridge the gap by automating tasks and aiding medical professionals in making quicker decisions
2. Data availability & analysis: the healthcare sector generates vast amounts of data that can be utilised to improve diagnoses, identify trends, and enhance patient outcomes. AI’s ability to process this data efficiently and provide insights is a significant advantage, especially in areas like medical imaging and diagnostics
3. AI’s proven capabilities: AI has already shown promise in areas like reading medical scans, processing biopsies, and detecting anomalies. There is strong investment and demand for such technologies to automate time-consuming tasks, improve diagnosis speed, and assist doctors in making informed decisions
4. Cost efficiency: the financial sustainability of healthcare systems, especially in the UK, is a concern. AI can lower costs by reducing administrative burdens, automating repetitive tasks, and speeding up diagnoses, thus improving the overall efficiency of healthcare delivery
5. Faster diagnosis & treatment: AI can drastically reduce diagnosis times, allowing for quicker identification of diseases and faster treatment. This would be especially beneficial in high-pressure environments like emergency care and oncology
6. Supporting overstretched professionals: there is a shortage of healthcare workers, and many professionals are overburdened. AI can alleviate some of this pressure by automating repetitive tasks, analysing large datasets, and even providing clinical decision support, reducing burnout and workload on staff
7. Advancements in AI tools: with the development of AI models tailored for healthcare, AI is increasingly being seen as a valuable partner to doctors and other healthcare providers. AI tools for early intervention, predictive analytics, and personalised care are rapidly advancing, offering new ways to improve patient care
8. Transforming patient care: AI not only enhances diagnostics but also allows for more personalized patient care. By analysing individual health data, AI can help customise treatment plans, monitor patient progress, and recommend interventions tailored to the specific needs of patients
9. Regulatory & ethical considerations: with appropriate governance, AI has the potential to transform the healthcare system, ensuring that its use is ethical and beneficial to all stakeholders. The alignment of AI with industry needs is driving its rapid adoption, as long as privacy and data security concerns are addressed
10. Machine learning & predictions: AI’s ability to recognise patterns in large datasets is vital for identifying health risks, predicting disease outbreaks, and offering preventive care. These predictive capabilities are particularly important in areas like genomics, drug discovery and early disease detection.
11. Public health surveillance: AI can help monitor and forecast health trends, improving the effectiveness of public health initiatives. By analysing vast quantities of health data, AI can support decision-making in disease prevention and resource allocation
12. Reducing bureaucracy: AI is also seen as a solution to reduce bureaucracy in the healthcare system. By automating administrative tasks, AI can remove redundant work, streamline processes, and reduce delays, ultimately enhancing service delivery

The verbatims for this area were plentiful — starting from the point that healthcare is already data-rich, and AI’s efficacy in pattern detection has already been proven in disease identification.

There are also societal issues at play here, as one member says: ‘due to the fractious nature of healthcare in the UK there are many opportunities for small, incremental, and immediate benefits to using GenAI for support. LLMs to enhance a level of trusted collaboration must be built for this to happen, and it is happening.’

And the end-user (the patient) is paramount here: ‘the increasing amount of data available that could be used to help train AI to assist healthcare professionals with decisions and management of patients could lead to better outcomes for patients. Patients need to be empowered and encouraged to be actively managing their interactions with healthcare providers.’

One member drew attention to the shortage of healthcare professionals across the world and their relative expensiveness to employ: ‘Both of these factors place a strain on the public and private sectors. Many aspects of healthcare involve analysis of data, which is an ideal use-case for AI. There is also a very large amount of training data available to aid in the creation of AI models for healthcare. All of these factors would make AI a very attractive prospect for healthcare organisations.’

Other comments covered the importance of the underpinning Google accessible information standard (AIS), and one posited that ‘health analytics is ahead of other sectors, with traditional blockers around identity/access and privacy/data protection being addressed at pace.’

Customer services

1. Cost savings: AI offers businesses the greatest option for reducing costs, especially in customer service. It is seen as a cheaper alternative to human agents, particularly for tasks like handling simple customer queries, admin work, and basic support
2. Rise of autonomous AI agents: the use of AI agents (e.g., chatbots) is already widespread and will continue to rise. These agents will handle low-level customer queries without human intervention, making services faster and more efficient, reducing wait times, and increasing customer satisfaction
3. Improvement in personalisation and scale: AI systems can be trained with company-specific data to provide personalised customer service, which can be scaled effectively. This is particularly important for industries with high customer volume and the need for quick, tailored responses
4. Automation of processes: many tasks currently handled by customer service representatives, such as ticket management, information retrieval, and responses to frequently asked questions, will be automated by AI. This not only streamlines processes, but also frees up human agents for more complex issues
5. Job reductions and impact on workforce: as AI takes over more customer service functions, there will be job cuts and redundancies in areas that involve repetitive or low-level tasks. However, this will also allow human employees to focus on higher-level, more complex service issues
6. Customer experience enhancement: by speeding up interactions and offering 24/7 availability, AI will likely improve the customer experience. AI tools such as chatbots can deliver faster responses, reduce frustration, and help customers self-serve, particularly for routine inquiries
7. AI's cost efficiency: with its ability to process large datasets and interact with customers at scale, AI becomes a cost-effective solution for companies looking to cut down on staff expenses while improving service delivery
8. Increased demand for AI solutions: AI in customer service is seen as a growing field, with businesses increasingly relying on AI to replace frontline agents or enhance their customer service operations. It’s expected that the shift will continue as AI tools become more sophisticated
9. Self-service and user preference: many customers prefer digital interactions over in-person communication, and AI provides a reliable and efficient method for self-service, further driving the shift toward automation

A recurrent theme in the verbatim answers to this question was that AI is already well-used in this space. A minority were somewhat cynical (LLMs will probably just ruin customer service,’ wrote one member), but many were not, even going so far as to say this is the most compelling use case.

As one commenter said: ‘Use of AI elements in CS chat functions are already widely adopted by businesses with an online presence. As software develops I believe it will likely replace a huge majority of customer service staff for all but the most specialist roles/requirements.’

Why? As one member put it, customer service ‘has the largest well-annotated body of data to train on, decades of historical ticket logs. It's such an easy win.. low-hanging fruit.’

One interesting possible issue was identified by this member: ‘AI can put large amounts of information at people’s fingertips. The caveat is that it should be the customer service advisors that get the AI support to help them; it's not going to be as effective if only put out as a public tool as not everyone knows how to ask questions in the right way to get the information out of chatbots.’

Developing AI skills

We asked members how they are considering developing their AI skills during the coming year. Here are the overall answers:

Protecting the public

With an eye on BCS’ broader societal concerns, we wanted to get a sense from our expert members on how we could best protect the public from AI’s unintended consequences. Here are the top answers:

In the verbatim comments, we had a variety of ideas. Governance was a key issue, alongside developing technical mitigations, and digital and data skills frameworks. ‘Governance frameworks that are focused on good practices,’ as one commenter said. ‘Regulation needs sufficient enforcement power and related funding, including a clear approach for reparations.’

Another commented on the need for ‘transparency and visibility of all data flows and components including where, by whom and how data (both personal and impersonal) is being used for training ML/AI.’

Whilst some felt that AI will not be as impactful as people currently think, there were a number of comments on how we get the balance between risk-taking and safety, with some advocating a slow-down to ensure we have a ‘comprehensive understanding of AI - including the learning models - currently we rush something out, and assume that it works generically as it works in the tested environments.’

A final comment on this: ‘We must embrace it (AI), and teach responsibility. You cannot legislate against the ocean tides. Strong social morality messages need planting now for a generation ahead - akin to the effective reframing of drink-driving - which was less of a sin when I was a child, but is now widely understood to be unacceptable. The message needs to start now in a structured way.’