Choosing digital tools in the age of AI

Marek Tuszynski – originally written for the Replaybook, September 2025, in section transforming: challenge tech paradigms; tactic 18 – Choosing Tools

Digital tools have become so embedded in the way we work that we almost forget that they are a choice. These choices increasingly matter for climate work as the consequences of these tools are becoming clearer. Their effects on our planet are already manifesting, but so too are their often-overlooked impacts on the way we think, interact, and engage with the world around us. This tactic explores how you can be more intentional about the tools you choose and use to support your work.

If you work on global climate issues, digital tools are likely essential to some part of your work. Yet the way you can make choices about using these tools - when, why, how, and which ones - is often messy and confusing. Digital tools aren’t like traditional tools. Historically, a tool was something tangible, something you held to build or fix things. That definition no longer fits. Many digital tools are intangible, complex, and constantly changing. Some, like smartphones, bundle multiple tools into one. Others act as gateways to larger systems - think of ChatGPT or Google Search.

Some tools evolve so rapidly that every time you use them, they may look, behave, or even be owned differently.

Most of the tools we use these days are not only cloud-based but were also created to be as simple as possible (frictionless) and to attract the largest possible number of users (social), who would in return become dependent on their platforms (user-based). These platforms are offered for free (you are the product) and they turn their user base into a data base (profiling) that can then be monetised (through targeted advertising or similar). We click ‘I agree’ to give them access to our behaviour (metadata) and facilitate cookies, beacons, scripts and what-not - anything goes, including fingerprinting browsers (trackers).

Another thought worth bringing up here is about dependency. Any collapse or disruption to these centralised services used by billions of people, companies, and even governments, would cause a problem of equal proportions. Large-scale dependency creates large-scale liability. We have already seen glimpses of what this can mean when services like Facebook/ WhatsApp or Google Docs go down for short periods of time. What if they went down for weeks? These are tools like any other – they have their limits and their critical mass. The question is when they are going to reach their limits and with what consequences. It is not about if but when.

These tools are more than just conveniences; they influence how we communicate, understand information, and engage with the world, while also exposing us, our peers, and our partners to various risks, which can sometimes be grave. This tactic examines the issues you need to consider when choosing the tools you work with and why.

Assess the costs

When considering the tools we want to use, we must recognise their explicit and hidden purposes, the risks they pose, and the costs of using them. These costs span, just to name a few:

Economic: Who pays? Who profits? Who is left behind?
Labour: Who builds and maintains them? Under what conditions?
Cultural/Social: What norms and values do they promote or erase?
Political: What agendas do they support or suppress?
Ecological: What resources - water, air, energy, land, minerals – do they consume or damage?

All of these risks and costs matter; it’s a case of assessing which ones you can manage and mitigate through your choices, and which are the most relevant for your organisation.

“These tools are more than just conveniences; they influence how we communicate, understand information, and engage with the world”Marek Tuszynski

Amidst the hype surrounding AI, we have become aware of the true ecological costs of digital technologies and how they have led to some tech companies abandoning or scaling back from their environmental goals (footnote 2). This includes acquiring power from nuclear plants(footnote 3) and using it to power data centres, and in some cases, this includes natural gas generators polluting the environment with methane(footnote 4). On this trajectory, the current growth of Big Tech is at odds with the increasing need to reduce our consumption of fossil fuels. This creates a significant dilemma for environmental groups who have been told that it is only a temporary problem, as renewables will solve it. However, given the pace of power-hungry AI, this currently seems more like wishful thinking (footenote5).

Assessing the risks: where to start

What new non-digital risks might emerge when digital tools are prioritised?

While successfully implementing a comprehensive set of digital tools can be perceived as progress, focusing too heavily on digital solutions can inadvertently introduce new risks in other areas. A holistic approach is essential when planning, implementing, or scaling up your digital infrastructure.

Key questions to consider:

What physical, emotional, or social risks might increase if digital security is improved? For instance, tighter digital controls could lead to threats in the physical world, such as increased surveillance or intimidation, or cause psychosocial stress within your team or among your partners.
Are digital tools being chosen based on a comprehensive risk assessment that considers more than just technical criteria?
Have you considered legal, cultural, political, and environmental risks, or only digital ones? Some vulnerabilities may only become apparent once a tool is in use.
Could implementing new tools create pressure points within your team or community?
Does the added responsibility, such as managing secure communications or learning complex systems, lead to an increased workload, stress, or burnout?
What non-digital dependencies are being created?
Does the new tool require constant access to electricity, the internet, or specific hardware? What would happen in areas where these are unreliable?
Are marginalised voices being excluded from the process?
Does a digital-first approach leave behind people who lack access, connectivity, language support, or technical literacy?
Does digital implementation shift risk to more vulnerable actors?

For example, could partners in repressive environments be at greater risk as a result of your digital transformation? What if the tool or its components are forbidden or illegal in some places?

Perhaps the most overlooked cost is the epistemological cost — the way in which these digital tools shape our understanding of the problems we face and the solutions we envisage.

For example, might generative AI, because of the data it was trained on and the agendas of those who own it, instead lock us in a bubble of indifference, preventing us from expanding our understanding of risks and solutions? After all, what we consider to be artificial intelligence is, in fact, corporate software as a service, presented to us as capable of acquiring and producing knowledge. It is often used instead of search engines or to aid the production of creative outputs. In fact, the software is exclusively capable of reproducing patterns (patterns of words, images, and code) based on the repositories of previous patterns it was trained on. However, can a pattern generation tool, however seductive, be considered the right tool to solve current and future problems(footnote 6)?

All of these factors need to play into the decision-making process around selecting what tools are right for you and your organisation and the task at hand.

Identify what matters to you

The way in which digital tools are approached depends on factors such as knowledge, resources, context, and skills. Each approach has its own advantages and disadvantages. As digital tools are constantly evolving, there is no single, rigid method that works for everyone.

The real questions, then, are: what kind of approach can we take towards something so ephemeral and complex, and why does it matter?

At Tactical Tech, for instance, we tend to favour open-source tools over proprietary ones, well-established tools over untested ones, and tools that demonstrate privacy protection in their actual code rather than just in marketing claims. But our evaluation doesn’t stop there.

We also consider ethical, political, and cultural factors, as well as the dependencies that a tool might introduce to our work. We ask: how will this change the way we operate? What new risks might it introduce while mitigating others (footnote 7)?

We consider the implications of scaling up: will the tool require additional technical expertise, server infrastructure, or security measures? Most importantly, we ask whether the tool truly serves our purpose. Will it strengthen us in one area while exposing vulnerabilities in another (footnote 8)?

Every organisation — and every individual — is different. We all have unique priorities, constraints, and value systems. Some of us have skills that others lack, and vice versa. This is why choosing digital tools is never just a technical decision: it’s always strategic, it is always tactical.

Ask the right questions

The first myth to dispel is the idea that you can simply download a tool that will solve all your problems and permanently align with your values. If that’s what you’re expecting, you’re not looking for a tool — you’re looking for snake oil. Digital tools are never standalone solutions. They’re often complex, unpredictable, and inconsistent. Some days they work reliably; other days, they don’t. They frequently over-promise and under-deliver. What seems effective today might become a liability tomorrow.

“Users routinely exchange personal data and behavioural insights for convenience and functionality. However, this exchange often creates long-lasting dependencies.”Marek Tuszynski

The checklist at the end of this chapter is the matrix we use to make our choices. The problem is that it’s difficult to find a tool that meets all your requirements, even the most essential ones. This matrix can be used as a framework to support decisions about adopting digital tools in a thoughtful, strategic, and values-aligned way. It is rare for a tool to satisfy all these criteria, but this structure helps you to weigh up the trade-offs clearly. Use it as a starting point and adapt it as necessary. When you have decided which tool you want to use, when, why, and how, then you may need some ideas of how to find tools other than the main ones everyone tends to use. There are alternatives you just need to know where to look, I explore this further in this piece: ‘The persistent problems of digital resilience’ (footnote 9).

As digital tools become increasingly central to how organisations and communities function, it’s important to approach their use thoughtfully and critically, especially when they’re used to address environmental and climate challenges. This reflection should not be the responsibility of just one department. While financial and technical assessments are important, we also need to consider the broader political implications and environmental costs of these tools. We should regularly ask ourselves whether these are the right tools for the challenges we face. If we are unsure, it may be worth seeking input from a wider range of voices, not just those within the tech industry.

Challenge tech paradigms

Choosing a digital tool is never just a technical decision — let me repeat it, it’s inherently strategic and tactical. Your choice will have consequences for your work, your team, your processes, and your long-term impact. At its core, the digital tool we casually refer to as an ‘app’ is actually a sophisticated socio-technical intervention that embeds specific ideologies, business models, and value systems. Beneath its sleek interface lies a web of tradeoffs: users routinely exchange personal data and behavioural insights for convenience and functionality. However, this exchange often creates long-lasting dependencies on not only proprietary platforms, providers, and hardware ecosystems, but also particular skill sets, licensing structures, and ways of thinking about the world.

Choosing the right tools; Asking the right questions

1. Transparency & Accountability

Can you trust the tool, and is it open to scrutiny? Who is behind it?

Is the tool open source, or is it a proprietary ‘black box’?
Is the codebase publicly available and regularly maintained?
Are there any controversies or known issues surrounding the tool or its developers?
Has the tool been independently audited or reviewed?
Is it clear who owns and funds the tool?
What would happen if the developers stopped supporting it, or if funding or legal conditions changed?
What type of license does the tool use, and who does it primarily protect: the user, the developer, or the tool itself?

2. Privacy & Security

How effectively does the tool handle data, protect your systems, and secure communication?

How clearly does it explain its approach to data privacy, communication security, and process protection?
Are its privacy policies transparent, specific, and enforceable?
What kind of encryption or other security measures does it use?
Is the tool based on widely accepted and tested security standards, or is it still experimental?

3. Ethical & Political Alignment

Does the tool align with your values and those of your partners and communities?

Does the tool’s development or usage align with your political, ethical, and social values?
Can you identify the positions or affiliations of its owners or maintainers?
Would using this tool compromise your principles, for example if the developers are known for discriminatory or unethical practices?

4. Cost & Resource Requirements

What will it cost in terms of finances, operations, and human resources?

Is the tool affordable within the current and future budget?
What are the initial costs (licensing, hardware, setup)?
What are the ongoing costs (maintenance, subscriptions, upgrades)?
What infrastructure, expertise, or staff will be required to use and maintain it?
Will you need to train the existing team or hire new people?
Can you realistically support that?

5. Usability & Fit for Purpose

Will the tool meet your needs and fit into your working methods?

Does the tool genuinely solve the problem at hand, or does it only address the symptoms?
Is this tool being considered as a shortcut to avoid addressing a deeper issue?
How well does the tool integrate with your current systems, workflows, and practices?
Will adopting it require significant changes to your infrastructure or habits?
Does the tool enforce a specific way of working that could limit your flexibility?
Is it available in the languages, formats, and accessibility levels required by you and your partners?

6. Support & Longevity

Can you rely on the tool in the long term, and will support be available when needed?

What kind of support is available for users? Is it responsive and accessible?
Is the documentation clear, up to date, and written in a user-friendly way?
Is there an active user or developer community around the tool?
Is the tool well-established and built on reliable standards?
What risks are involved in using this tool, in terms of the technical, organisational, and reputational aspects?

7. Risk, Resilience & Impact

What risks are involved in adopting this tool, and how would your team cope if something went wrong?

How resilient is your work if this tool becomes unavailable or fails?
Could this tool compromise the safety or privacy of your team or partners?
What is the tool maker’s track record when it comes to acknowledging and mitigating their environmental impact?

8. Environmental & Social Impact

What are the broader environmental and human costs of using the tool?

What environmental resources does the tool rely on, such as energy, water, or minerals?
Are there any indirect social costs associated with its development or operation, such as labour exploitation or unsustainable supply chains?
How does the tool’s use align with your environmental goals and responsibilities?

“The choice of digital tools is never neutral. it requires reflection, responsibility, and an honest assessment of values and risks. and this reflection should not be the responsibility of just one department or the tech-person in your organisation - it requires collective understanding. while this framework won’t provide easy answers, it will help you ask the right questions”Marek Tuszynski

Footnotes

1. This material is taken verbatim from the author’s own work, previously self-published, and now available at: Tuszynski, M., ‘Technology is stupid’, Tactical Tech (April 2020)
2. Milmo, D., Hern, A., and Ambrose, J., ‘How AI’s insatiable energy demands jeopardize Big Tech’s climate goals’, Mother Jones (July 2024); ‘AI could keep us dependent on natural gas for decades to come’, MIT Technology Review (May 2025)
3. Lawson, A. ‘Google to buy nuclear power for AI datacentres in “world first” deal’, The Guardian (October 2024)
4. Wittenberg, A., ‘“How come I can’t breathe?”: Musk’s data company draws a backlash in Memphis’, Politico (June 2025)
5. ‘AI is poised to drive 160% increase in data center power demand’, Goldman Sachs (May 2024)
6. Bender, E.M., Gebru, T., McMillan-Major, A., and Shmitchell, S., ‘On the dangers of Stochastic Parrots: Can Language Models be too big? In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (FaccT ‘21)’, Association for Computing Machinery, New York (2021)
7. Hankey, S., and Tuszynski, M., ‘Efficiency and madness: Using data and technology to solve social, environmental and political problems’, Heinrich Böll Foundation (2017)
8. Tuszynski, M., ‘Technology is stupid’, Tactical Tech (April 2020)
9. Tuszynski, M., ‘The persistent problems of digital resilience’, Tactical Tech (January 2025)