Table of contents

Approximate read time: 13 minutes

The House of Lords is scheduled to debate the following motion on 5 June 2026:

The Archbishop of Canterbury to move that this House takes note of the impact of artificial intelligence on human relationships and society.

1. Key points

  • Over half of adults aged 16 and over use generative AI, with the highest usage rates among the youngest cohorts.
  • Some people use AI for companionship; though this was the least common use in the UK in 2025, it was the most common use in the US.
  • Evidence of the impact of using AI on human relationships is limited. Some posit that increased use of AI for companionship could reduce incentives to maintain human relationships, with negative long-term social consequences.
  • There are also potential benefits of AI to support vulnerable and neurodivergent users in their human relationships.
  • AI’s wider societal impact is likely to be significant, but evidence is still emerging. Implications are expected for work, education, healthcare, the creative industries and the environment, among other areas.
  • It is thought some sectors may see major change with productivity gains and job disruption in the labour market possible. There are also developing legal and ethical challenges, such as in copyright.

2. Artificial intelligence and human relationships

2.1 Growth of conversational artificial intelligence

Recent years have seen rapid advances in artificial intelligence (AI). Many people now use generative AI tools that can engage in human-like dialogue. In 2025 just over half of over-16s in the UK used generative AI tools such as ChatGPT, Copilot and Gemini.[1] Use was highest among the youngest cohorts, with 79% of 16–24-year-olds and 74% of 25–34-year-olds using these tools.

The most common uses of generative AI were for work or study or finding out factual information.[2] However, 12% of generative AI-users reported that they used it as a friend or as someone to talk to.

Table 1. Reasons for using generative AI, UK over-16s
Use Percentage of generative AI-users who use AI this way
Work or educational purposes 47%
Factual information 45%
Curiosity/to find out more about it 43%
Writing/editing tasks 39%
Fun or creative tasks 38%
Part of job/work related 33%
Recommendations 31%
Part of studies/education 26%
Someone to talk to/as a friend 12%

(Ofcom, ‘Adults’ media use and attitudes report’, 2 April 2026, p 11)

A separate study in the US found that therapy and companionship was the number one use of generative AI in 2025.[3]

Some AI applications are specifically designed to provide companionship, including romantic relationships. Some of the biggest companies providing this service include Replika, Character.ai, PolyBuzz and Chai.[4] These applications allow users to anthropomorphise the AI, choosing its appearance, gender, accent and other characteristics. The AI retains and builds on previous interactions, allowing the user to feel they are developing a relationship with the AI.[5]

2.2 Possible impact on human relationships

As widespread use of generative AI is relatively new, evidence of its impact on human relationships is limited. However, some academics and commentators have posited ways in which using AI for companionship could affect human relationships.

One theme in the discourse is that using AI for companionship could negatively affect human relationships because AI companions do not require reciprocity in the same way humans do; AI companions are always available, and do not have their own needs.[6] Therefore, people may find interacting with an AI companion easier than sustaining a human relationship. Writing in the journal AI and Society, Kim Malfacini of Open AI summarises this theory: “As companion AI learns to meet our needs more, we learn to meet each others’ less”.[7] Academics Justin Keeler and Brett Murphy argue that as a result, “it is possible to imagine a future where, with one or a few companion chatbots, some people may see little need for the messiness and effortfulness of human relationships”.[8]

Professor Rahul Ravi frames this problem as people prioritising short-term over long-term rewards.[9] AI companions give immediate, short-term rewards and do not require very much work. In contrast, human relationships require time and “a willingness to tolerate discomfort”. Professor Ravi argues that “by reducing effort, uncertainty and emotional risk, AI companions make connection easier to access but may also shift expectations in ways that are harder to sustain over time in human relationships”.

Relatedly, some have argued that using AI companions could lead to people becoming less skilled at interacting with humans, particularly if used in childhood and adolescence.[10] However, others have argued that AI has the potential to improve people’s interpersonal skills if it is designed to do so. AIs could be designed to “encourage users to treat them with care and consideration”.[11] In addition, AI companion apps can “offer a safe space for users to rehearse social interactions”.[12] Kim Malfacini highlights that opportunities to practice social skills with AI can be useful for some people who are neurodivergent.[13]

3. Other societal implications

The potential impact of AI on society generally is extremely broad. The section below provides a brief explanation of the current and potential impact of AI in some areas and a selection of further reading on each topic.

3.1 Work

The impact of AI on work and how this impact will be distributed is emerging but remains uncertain.[14] For example, some evidence from the UK and US suggests employment is already declining in some jobs heavily exposed to AI, such as computer programming. However, employment has also declined in other sectors not greatly exposed to AI, such as hospitality, and the impact of other factors such as interest rates and business cycles cannot be ruled out.

The Office for Budget Responsibility (OBR) has estimated that “AI could materially impact 40 percent of the UK labour force over the next 10 years, with the majority of occupations expected to be complemented by AI rather than substituted”.[15] It finds that occupations composed largely of tasks that could be automated, such as administrative, secretarial, sales and customer service roles, are the most exposed. The least exposed are those mainly consisting of physical tasks, such as cleaners, skilled trades, and process, plant and machine occupations. The OBR concluded that AI would be likely to increase annual productivity growth by up to 0.8 percentage points over the next decade.[16]

Read more about AI and work

3.2 Education

A 2025 government report found at present there is “limited evidence on the impact of AI use in education on learners’ development, the relationship of AI use and educational outcomes, and the safety implications of children and young people using this technology in the classroom”.[17] Generative AI is suited to many educational tasks, such as creating resources, lesson planning and administrative tasks, and could free up teachers’ time to focus on interacting with students. However, content generated by AI can also be inaccurate, inappropriate or biased.

An April 2026 National Education Union survey collected teachers’ views on and experiences of AI in education. It found:[18]

  • Two-thirds of secondary teachers think that pupils’ critical thinking has declined due to AI usage. This is more than double those working in primary.
  • AI use among teachers is now widespread and growing. Three quarters are now using AI tools for day-to-day work, up from 53 percent the previous year.
  • Teachers primarily use AI for resource creation but also lesson planning and administrative tasks. Seven percent use AI tools for marking.
  • Half of schools have no policy on the use of AI either by staff or students. Two-thirds have no policy specifically for students.

Read more about AI and education

3.3 Health

AI has a range of potential applications in healthcare, both administrative and clinical.[19] These systems can fully automate processes or be used as tools under human supervision with the aim of reducing costs and improving health outcomes.

The government’s 10-year health plan for England seeks to “make the NHS the most AI-enabled health system in the world”.[20] AI is already being used in some capacities in the NHS.[21] For example, a recent trial involving over 30,000 NHS staff reportedly saved each staff member 43 minutes a day on average by automating administrative tasks.[22] Current clinical applications include identifying disorders from blood samples and detecting early signs of lung cancer in X-ray and CT scan images.[23] Work on a new NHS-wide cloud computer system is ongoing to further enable the use of AI-screening technologies across the country.[24]

Some commentators in healthcare have argued that AI will be essential for the NHS to meet rising demand, although others argue practical, regulatory and ethical challenges may limit its use.[25]

Read more about AI and health

3.4 Creative industries

There are two major themes at the intersection of AI and the creative industries. One is how generative AI can complement, or replace, creative work. Another is the question of whether companies can legally train their AI models on copyrighted material.

As in other sectors, there is emerging evidence that AI can increase productivity in the creative industries but also concern that it could lead to fewer jobs in some sub-sectors.[26] Some occupations, such as composers and animators, are highly exposed to AI. Others, such as dancers, are much less exposed. So far there is little discernible impact of AI on earnings in the more exposed occupations: in the US in 2024, earnings trends for artistic occupations with higher exposure were broadly similar to those with less exposure.[27] Some more exposed occupations experienced weaker employment growth in 2023 than less exposed ones, but the differences were small. However, UNESCO, the UN’s cultural organisation, predicts that generative AI is likely to result in significant revenue loss for both music and audiovisual creators by 2028.[28]

Generative AI systems are developed using text and data mining techniques that involve the copying and analysis of large quantities of data to identify statistical patterns and relationships.[29] Much of the material used in developing and operating generative AI systems consists of copyright-protected works such as books, news articles, computer programs, photographs, music and films. While there are points in the process of training AI models where copyright may be engaged, this is disputed.[30] The government has characterised the current position as legal uncertainty, but creative industry professionals have argued that using copyright-protected works to train AI models is an infringement. In 2025, consultees largely rejected the government’s proposed reforms to copyright law in light of AI.[31] Following this, the government said it did not have a “preferred option” on how to address the issue and would not introduce reforms to copyright law unless it was confident they would meet its “objectives for the economy and for UK citizens”. It has outlined other actions it is taking in the area, including:[32]

  • launching a consultation on digital replicas
  • establishing a taskforce on AI labelling
  • reviewing the mechanisms for creators to control their works online
  • launching a working group on independent and smaller creative organisations

Read more about AI and the creative industries

3.5 Environment

The increase in use of AI is predicted to have a significant environmental impact.[33] Environmental impacts of AI include:[34]

  • The computational power required to train generative AI models requires large amounts of electricity, leading to increased carbon emissions if this comes from non-renewable sources.
  • Each use of an AI tool also requires a significant amount of energy; a ChatGPT query uses approximately five times more energy than a simple web search.
  • Large amounts of water are needed for cooling generative AI’s hardware.
  • The hardware used in generative AI requires natural resources for its manufacture and transport.

The energy consumption of data centres globally has increased significantly since 2022, partly driven by development of generative AI.[35] It is expected that by the end of 2026 data centres will consume 1,050 terawatt hours, equal to an amount between the annual electricity consumption of Japan and Russia.

In addition, the mining of the raw materials needed to make the processors used for AI often involves unsustainable practices.

Read more about AI and the environment

Image by Kohji Asakawa from Pixabay.

References

  1. Ofcom, ‘Adults’ media use and attitudes report’, 2 April 2026, p 9. Return to text
  2. As above, p 11. Return to text
  3. Harvard Business Review, ‘Data and visuals: Top 10 gen AI use cases’, 9 April 2025. Return to text
  4. TechCrunch, ‘AI companion apps on track to pull in $120mn in 2025’, 12 August 2025. Return to text
  5. American Psychological Association, ‘AI chatbots and digital companions are reshaping emotional connection’, 1 January 2026. Return to text
  6. Kim Malfacini, ‘The impacts of companion AI on human relationships: Risks, benefits, and design considerations’, AI and Society, 16 April 2025, vol 40, pp 5527–40; and Justin B Keeler and Brett A Murphy, ‘Chatbots and human-human relationships: The need for research on potential downstream harms from generative AI’, Community, Work and Family, 6 February 2026. Return to text
  7. Kim Malfacini, ‘The impacts of companion AI on human relationships: Risks, benefits, and design considerations’, AI and Society, 16 April 2025, vol 40, pp 5527–40. Return to text
  8. Justin B Keeler and Brett A Murphy, ‘Chatbots and human-human relationships: The need for research on potential downstream harms from generative AI’, Community, Work and Family, 6 February 2026. Return to text
  9. Rahul Ravi, ‘From AI companions to climate action, we undervalue what lies ahead’, The Conversation, 11 May 2026. Return to text
  10. Justin B Keeler and Brett A Murphy, ‘Chatbots and human-human relationships: The need for research on potential downstream harms from generative AI’, Community, Work and Family, 6 February 2026. Return to text
  11. As above. Return to text
  12. American Psychological Association, ‘AI chatbots and digital companions are reshaping emotional connection’, 1 January 2026. Return to text
  13. Kim Malfacini, ‘The impacts of companion AI on human relationships: Risks, benefits, and design considerations’, AI and Society, 16 April 2025, vol 40, pp 5530. Return to text
  14. Department for Science, Innovation and Technology, ‘Assessment of AI capabilities and the impact on the UK labour market’, 28 January 2026. Return to text
  15. Office for Budget Responsibility, ‘Briefing paper no 9: Forecasting productivity’, November 2025, p 47. Return to text
  16. As above, p 52. Return to text
  17. Department for Education, ‘Generative artificial intelligence (AI) in education’, updated 12 August 2025. Return to text
  18. National Education Union, ‘State of education: AI’, 2 April 2026. Return to text
  19. World Health Organisation, ‘Ethics and governance of artificial intelligence for health: Guidance on large multi-modal models’, 25 March 2025. Return to text
  20. Department of Health and Social Care, ‘10 year health plan for England: Fit for the future’, 3 July 2025. Return to text
  21. Barts Health NHS Trust, ‘AI already transforming hospital care’, 31 March 2025. Return to text
  22. Department of Health and Social Care and NHS England, ‘Major NHS AI trial delivers unprecedented time and cost savings’, 21 October 2025. Return to text
  23. Barts Life Sciences, ‘How the NHS and academia collaborate to speed up diagnosis for blood disorders’, May 2025; and Peninsula Imaging Network, ‘New AI tool is helping in the treatment of lung cancer’, 11 April 2025. Return to text
  24. Department of Health and Social Care and NHS England, ‘AI to be trialled at unprecedented scale across NHS screening’, 22 September 2025. Return to text
  25. Royal College of Radiologists, ‘AI deployment in the NHS: Reviewing progress made and defining future action’, April 2025; Jessica Morley, ‘AI and the NHS: Is it the silver bullet that will improve the health service’s productivity?’, Nuffield Trust, 15 August 2024; and Malte Gerhold, ‘10-year health plan: We need to move from ‘techno optimism’ to ‘techno realism’’, Health Foundation, 17 July 2025. Return to text
  26. London School of Economics Growth Lab, ‘AI and the creative industries’, June 2025, pp 5–6 and 10–11. Return to text
  27. Gallup, ‘AI is changing creative work, but the arts aren't disappearing’, 4 May 2026. Return to text
  28. UNESCO, ‘Reshaping policies for creativity’, 2026, p 64. Return to text
  29. House of Lords Communications and Digital Committee, ‘AI, copyright and the creative industries’, 6 March 2026, HL Paper 267 of session 2024–26, p 10. Return to text
  30. As above, p 16. Return to text
  31. House of Lords Communications and Digital Committee, ‘Government response to the House of Lords Communications and Digital Committee’s report on ‘AI, copyright and the creative industries’’, 15 May 2026, p 2. Return to text
  32. As above, p 1. Return to text
  33. Apoorva Chouksey et al, ‘The green paradox: The climate, environmental, and sustainability implications of artificial intelligence’, Global Environmental Change Advances, March 2026, vol 6. Return to text
  34. MIT News, ‘Explained: Generative AI’s environmental impact’, 17 January 2025. Return to text
  35. As above. Return to text