AI at the Bentway

These summaries draw on 18 staff responses. To protect privacy in a small team, this version includes no direct quotes or individual response records. Charts and counts include all responses to each question.

  • Drafting & editing writing

    8 of 18 staff
    Communications

    Staff use chatbots to help write and edit emails, copy and reports. Some described writing a first draft themselves, asking AI to improve it, then editing the result to sound like them.

    Open canvas →
  • Research & background gathering

    8 of 18 staff
    Research & Development

    Staff use AI to learn about unfamiliar subjects and find leads for further research. Responses also describe checking other sources, which can take as much time as researching without AI.

    Open canvas →
  • Data & spreadsheets

    8 of 18 staff
    Planning & analysis

    Staff use AI to compare contact lists, format spreadsheets and prepare data for presentations. Responses describe uploading files, checking the results and sometimes finding errors that take longer to fix than doing the work manually.

    Open canvas →
  • Admin & scheduling

    5 of 18 staff
    Scheduling & logistics

    Staff use AI to organize information and put together schedules and itineraries. These tasks can involve personal information, such as an artist’s travel or contact details.

    Open canvas →
  • Summarizing documents & meetings

    4 of 18 staff
    Document review

    Staff use AI to summarize long documents and meeting notes, and to turn feedback from several people into suggested edits. Summaries still need to be checked against the original material.

    Open canvas →
  • Brainstorming & ideas

    4 of 18 staff
    Arts & Culture

    Staff ask chatbots for ideas for grants, programs and installations. Responses describe using the suggestions to develop their own thinking and research.

    Open canvas →
  • Grants & funder reporting

    3 of 18 staff
    Fundraising

    Staff use AI to develop grant ideas, edit applications and reports, and adapt wording for different funders. One response described substantial time savings, alongside reluctance to tell external partners or funders about using AI.

    Open canvas →
  • Translation & identification

    2 of 18 staff
    Translation & Language Access

    Staff reported using AI for translation and image searches to identify plants or machine parts. One response described checking suggested plant names against other websites before acting on the results.

    Open canvas →

All staff, all uses

18 staff
How oftenIn a typical month, how often do you use AI tools for work?
32643
NeverRarelyA few times a monthA few times a weekDaily
Trust in own useWhen you use AI at work, how confident are you that you’re using it appropriately?15 answered
627
Not confidentRarely sureSometimes sureMostly sureAlways sure
Time saved / weekRoughly how much time do you estimate AI saves you in a typical week?9 answered
2511
NoneUnder 1 h1–3 h4–7 h7 h +
Comfort disclosingWould you feel comfortable telling a colleague or your manager that you used AI on a piece of work?15 answered
2310
Would hide itProbably notIt dependsProbablyOpenly yes

Each cell shows an answer choice. Darker cells mean more people chose that answer. The pointer below the bar marks the median.

What worries staff

  • Losing the human voice in our communications14/18
  • Accuracy of what it produces12/18
  • Environmental impact12/18
  • Privacy & confidentiality of what we put in11/18
  • How partners, funders or the public would react9/18
  • Effects on jobs or roles6/18
  • No significant concerns2/18

What staff want the AI policy to answer

  • Which uses of AI are acceptable, and which tools can staff use?
  • What information can staff safely share with AI tools, and what privacy and security protections are in place?
  • How should the Bentway account for AI use in its environmental impact reports?
  • What is the Bentway’s policy on AI-generated images in its own materials?
  • Could staff or job applicants be penalized for choosing not to use AI?
  • How should staff use AI to assist their work, and what review is expected?
  • Should public materials avoid AI-generated writing and images, or be edited so they do not look or sound AI-generated?
  • Do we tell partners and funders when AI helped with a piece of work?