Our Position on AI

Our Approach

• We operate independently from any single AI provider, focus on results and avoid hype.
• Editorial control stays with people. AI can be useful for drafting, but it doesn’t decide what we say. We don’t publish AI-only outputs as finished deliverables.
• AI is best used to enhance workflows, not to replace people doing skilled work.
• We disclose AI use to clients and seek consent. Where AI features in a project, it’s recorded in project documentation and surfaced through routine reporting.
• We use tools and settings that keep client or project data out of model training pools.
• AI’s unsustainable environmental impacts mean that, as with travel, its use must bring significant benefits to values-led work, with mitigations sought where possible.

We deploy AI with and for people, in line with this approach, carefully assessing benefits and trade-offs, including possible unforeseen impacts. Before we look at things the technology can help people achieve, aspects like reliability, appropriate data management and security protocols provide the foundation.

Where We Started

In 2017, we began using early machine learning to produce first-pass translations on international projects. Native speakers and professional translators checked and improved them before anything went out. Our approach hasn’t changed. AI is there to enhance good work and free up people’s time to focus on more meaningful, strategic and added-value tasks.

Worked Examples

AI features across our knowledge, project and communications management work, and we help clients and projects build their own AI-supported workflows in these areas. The examples below provide a few snapshots from recent projects:

How We Work With Clients on AI

Some clients want help thinking through where AI could responsibly fit in their own work. That tends to start with the boring but important things like data, consent, procurement, opt-outs, and what the technology is good at versus what’s currently hype. From there, we can scope use cases, deliver them, or pilot and embed them within teams. The areas are the same ones where we deploy AI ourselves: knowledge, project and communications management.

The Thinking Behind Our Position

The Big Picture

Unless some form of currently mythical ‘AGI’ (Artificial General Intelligence) is achieved and human oversight is abandoned, AI is likely to remain a powerful tool that enhances human work rather than fully displacing people’s judgment and expertise. AI is only as good as the people using it, and you have to know what good looks like from the start. Adoption is also making clear where human soft skills are irreplaceable.

The Current Situation

The technology’s uptake so far supports this view. For every administrative or technical role being ‘streamlined’ or ‘capacity-boosted’, the importance of human checks and accountability increases, especially around accuracy, security and liability.

The same applies to strategic and creative roles. Studies show AI is decent at sourcing and synthesis but poor at strategy, making high-quality knowledge management and thinking essential. In a world where simulated content is easy to generate, editorial judgement, authentic human stories, and creative expertise matter more than ever.

Some Context

We’ve been here before. While it’s not a perfect analogy, what’s happening today is similar to the rollout of computing, the internet and broader digital technologies in the 1990s and early 2000s. Taskscape was founded in 2006 partly in response to that earlier wave of transformation, on the view that without support, values-led initiatives would fall behind.

A Question of Values

While principled, ignoring AI isn’t the answer. Purely profit-driven outfits and individuals won’t, leaving purpose-led initiatives at a disadvantage. Whether the current investment bubble bursts or not, we suspect AI will shape some of the best and worst aspects of the coming decades.

Does AI Actually Work?

The answer depends on what your goals are and how implementation is carried out. Many organisations have been implementing AI while making a series of key mistakes:

• Focusing on trying to find efficiencies that reduce staff count and costs.
• Thinking higher volume or speed means higher net productivity by default.
• Falling victim to hype and not doing due diligence on reliability and results.
• Trialling or implementing tools without properly scoping the usage costs.
• Not understanding what AI tools are actually good at and best used for.

Most organisations making these errors are purely profit-driven, but anyone can fall foul of them. We know that, when deployed properly, AI can be a phenomenally useful tool, because we’ve done it for ourselves and clients. The good news is that solutions across knowledge, project, and communications management are among the most reliable, measurable, and beneficial in an extremely uneven landscape of tools.