Control in the Age of Continuous Information: Rethinking CAI in the Booth

For some time now, discussions around computer-assisted interpreting have revolved around a deceptively simple question: can these tools help us?

It is a natural question, and yet, Iincreasingly feel that it is the wrong one.

The issue is not whether a tool helps. The issue is whether, in using it, the interpreter remains in control.

This distinction may appear subtle at first glance. In practice, it defines the difference between a tool that integrates into professional performance and one that competes with it.

The Source of Discomfort

Many interpreters, including highly experienced colleagues, express reservations about CAI tools. These reservations are often framed in terms of distraction, cognitive overload, or a perceived degradation in output quality. One observes colleagues who seem tofollow captions too closely, whose delivery becomes more rigid, or whose attention appears divided between listening and reading.

These observations should not be dismissed. They are grounded in real experience. However, they are frequently interpreted as evidence that CAI itself is inherently problematic.

A more precise reading would be this: what is problematic is not the presence of additional information, but the way in which that information is introduced and handled.

The Assumption of Selectivity

A common response to this discomfort has been to advocate for selective display. If too much information is distracting,the reasoning goes, then tools should show less: keywords, numbers, or selected terms deemed relevant in advance.

This approach is intuitively appealing. It seeks to reduce visual load and preserve the primacy of listening. Yet it rests on an assumption that does not hold under real working conditions: namely, that we can reliably predict what will be relevant before it becomes relevant.

In practice, meaning often emerges retrospectively. A number that seemed secondary becomes central. A term that appeared familiar reveals a different nuance in context. A sentence that beginsin one direction resolves in another.

Selective display, by definition, filters information before its relevance is fully established. In doing so, it risks withholding precisely what the interpreter may need at a critical moment.

Continuous Information as a Different Paradigm

An alternative approach is to abandon the attempt to pre-select and instead provide continuous, structured information in the form of full text.

At first encounter, this may seem counter intuitive. Would this not increase cognitive load rather than reduce it?

The answer depends on how we understand the role of that information.

If full text is treated as something to be read continuously, then indeed it becomes intrusive. But this is not its intended function. Its value lies in its availability. It is there not to befollowed line by line, but to be accessed when required, without delay.

In other words, it does not ask for attention. It offers support at the moment attention turns to it.

Attention as a Trained Capacity

This brings us to the core of the matter. The challenge posed by continuous information is not primarily technological; it is cognitive.

Working with a persistent layer of text requires a form of selective attention that is not immediately intuitive. The interpreter must learn to allow information to remain in peripheral awareness without engaging with it unnecessarily, and to shift attention to it briefly and efficiently when needed.

This is, in essence, a skill.

It is worth recalling that simultaneous interpreting itself is not a natural activity. It is acquired through sustained practice, during which the interpreter learns to manage multiple processes in parallel: listening, analysis, reformulation, and production. What initially appears overwhelming becomes, over time, integrated and controlled.

The same principle applies here. Continuous textual information introduces a new layer, but it does not fundamentally alter the nature of the task. It extends it.

The Limits of “On-Demand” Visibility

Some have proposed that information should remain hidden until explicitly needed, appearing only upon request. While this approach reduces visual presence, it introduces a different constraint: the delay between need and access.

In real-time interpreting, such delays are not trivial. The moment at which a term, number, or segment becomes relevant is often the moment at which it must already be available.

Continuous presence, by contrast, ensures immediacy. The interpreter does not need to request information; it is already there, ready to be consulted if and when required.

The distinction is subtle but significant.It shifts the model from reactive retrieval to proactive availability.

Ergonomics and Control

None of this implies that all information should be imposed uniformly on the interpreter’s field of vision. On the contrary, control over how information is presented remains essential.

Interpreters must be able to configure their working environment: to adjust positioning, scale, and visibility; to reduce visual competition when necessary; and, at times, to remove the interface altogether.

These adjustments, however, concern the interface, not the underlying availability of information. The content remains intact. What changes is the interpreter’s relationship to it.

The Learning Curve

It is at this point that many evaluations of CAI tools become premature.

Initial encounters often produce discomfort. Attention feels divided, the presence of text intrusive. From this, it is tempting to conclude that the approach itself is flawed.

Yet this mirrors a familiar pattern. Early experiences of simultaneous interpreting are rarely comfortable. The coordination of listening and speaking under time pressure initially results in hesitation, errors, and fatigue. Only through repetition does coherence emerge.

CAI introduces a comparable adjustment. It is not a shortcut, nor an immediate enhancement. It is an additional competence that develops over time.

To dismiss it at an early stage is to judge the skill before it has had the opportunity to form.

Accessibility and Professional Development

This has implications beyond individual practice. If tools require sustained engagement to be used effectively, then accessibility becomes a central issue.

Barriers to entry, whether financial or technical, reduce the likelihood that interpreters will invest the time needed to adapt. Without widespread adoption, the collective learning process remains limited, and the profession does not fully explore the potential of these tools.

In this sense, the question is not onlywhether a tool is effective, but whether it is accessible enough to be meaningfully adopted.

Interpreting in an Evolving Landscape

The broader context cannot be ignored.Advances in AI-driven speech processing continue to reshape expectations around language mediation. While such developments are often framed in terms of replacement, they also highlight the distinctive nature of human interpreting.

Interpreting is not merely the transfer of words. It is the management of meaning in conditions of uncertainty, speed, and incompleteness. It involves prioritization, adaptation, and judgment, processes that are not easily reduced to algorithmic output.

Tools that align with this reality can reinforce professional capacity. Those that disrupt it risk undermining it.

A Subtle but Significant Shift

What is emerging, gradually, is a shift in how interpreters engage with information.

Rather than limiting input, the focus moves towards managing it. Rather than avoiding complexity, the emphasis is on maintaining clarity within it.

Those who adapt begin to exhibit adifferent relationship to the tool. They do not follow it; they consult it.They do not depend on it; they integrate it. The text becomes neither central nor irrelevant, but available.

This balance is not immediate. It is developed.

Conclusion

The question, then, is not whether CAI tools assist interpreting. It is whether they support the interpreter’s ability to remain in control in an environment where information is increasingly abundant.

Full-text, continuous information is not inherently a source of distraction. It becomes one only when the skills required to manage it are not yet in place.

As with simultaneous interpreting itself, these skills can be acquired.

The future of the profession will not be determined by the presence or absence of tools, but by how effectively interpreters learn to work with them.

Control, in this context, is not given.

It is trained.

‍