Interpreter

United States. The statistics bureau projects that employment of interpreters and translators, counted together, will grow 2 percent from 2025 to 2035, from 73,900 to 75,400, with about 6,000 openings a year. It says demand for American Sign Language interpreters is expected to grow due to the increasing use of video relay services, and that AI is making the work of translators and localization specialists more efficient; it gives no separate projection for interpreters.

A named person with standing publicly predicted something, on a date, in an attributable statement. It is recorded so that who said what, and when, stays checkable — and it never moves a task's assessment, because a prediction is not an observation. Its value arrives later: the record sits on the same page as the evidence about that occupation, so anyone reading the forecast reads the record of what happened next beside it. That is the reckoning; this site publishes no verdict on whether a forecast came true.

United States, a clinic setting, six languages. Using real clinical speech segments, the authors report that Google Translate in speech/voice mode had significantly higher linguistic and clinical error rates than qualified medical interpreters (33.3% vs. 4.8%), and that errors were worse for less commonly spoken languages (42.9% vs. 14.3%). It is a small exploratory study of one app.

A demo, benchmark or paper shows the task can be done. Updates what the technology can do — not what employers will do.

European Union institutions. The guidance says AI tools can help prepare interpretation work and interpreting in the booth, warns that they can pose a risk to sensitive information and personal data, and recommends the Commission's own eTranslation and Speech-to-Text tools instead of commercial solutions. It is an employer's guidance on its interpreters' own use of AI; it says nothing about replacing interpreters.

Measured, large-scale use of a tool for real work, where the decision to use it was the worker's rather than an employer's. It is more than a capability record — the work is real, not a demo — and less than a deployment record, because no employer put it into production, required it, or built a process around it. Weighted `cautious`: `automating` means the machine can do the task AND there are adoption signs, and this is an adoption sign — but usage can be experimental, and much of the measurement comes from a party with a stake, so one record is never enough and two independent ones are. Note who is counting. Vendor telemetry sees this directly and sells the tool, so such a record names that stake in its scope; a statistics agency asking firms whether their workers use AI in tasks sees the same channel with no stake at all, and that is the better source where it exists.

United States, state courts. The guide says courts should never use machine translation for court events, to convey legal or procedural information or to carry out complex, extended interactions, because errors could affect people's ability to understand and exercise their legal rights. It is guidance from a court-support organisation, not a binding rule.

Failure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.

A technology company's research paper on its own simultaneous speech translation system. The authors report that on real-world speeches, which are often disfluent, informal and unclear, CLASI achieves 81.3% and 78.0% on their valid information proportion metric for Chinese-to-English and English-to-Chinese. The metric and the parity threshold are the authors' own; it is not a head-to-head test in a real booth.

A demo, benchmark or paper shows the task can be done. Updates what the technology can do — not what employers will do.

United States, health programs receiving federal funds. The final rule says that when interpretation services are required, a covered entity must offer a qualified interpreter, and that if it uses machine translation when the underlying text is critical to an individual's rights, benefits or meaningful access, the translation must be reviewed by a qualified human translator; the rule defines machine translation in terms of text. Effective July 5, 2024. It is one country's rule; other parts of the rule have been challenged in court.

Regulation, subsidy or public procurement is requiring or funding adoption — the mirror of a constraint. It shows adoption is being required, not that it has happened, so one mandate is never enough on its own; two independent ones are.

United States. Using a non-inferiority design with professional medical interpreters as the gold standard, the authors evaluated three machine interpretation apps for English–Spanish and English–Mandarin brief clinical communication; Spanish or Mandarin to English machine interpretation had low acceptable scores (0.36 to 0.41), and no app met the non-inferiority threshold. The authors advise clinicians to defer use until the applications improve. It tested 2022-era apps on simple phrases.

A demo, benchmark or paper shows the task can be done. Updates what the technology can do — not what employers will do.

A comparison of human simultaneous interpreters and a machine simultaneous speech translation system on the same speeches. The authors found better performance for the human interpreters in terms of intelligibility, while the machine performs slightly better in terms of informativeness. The sample is small, and the first author is affiliated with a remote interpreting company.

A demo, benchmark or paper shows the task can be done. Updates what the technology can do — not what employers will do.

European Union. The directive says interpretation provided under it shall be of a quality sufficient to safeguard the fairness of the proceedings, in particular by ensuring that suspected or accused persons know the case against them and can exercise their right of defence, and that member states shall endeavour to establish registers of appropriately qualified independent interpreters and translators. It sets a quality standard, not a rule against machines; member states implement it in their own law.

Regulation, subsidy or public procurement is requiring or funding adoption — the mirror of a constraint. It shows adoption is being required, not that it has happened, so one mandate is never enough on its own; two independent ones are.

United States, state and local government services. The regulations define a qualified interpreter as an interpreter who, via a video remote interpreting service or an on-site appearance, is able to interpret effectively, accurately and impartially, and require video interpreting to show a sharply delineated image large enough to display the interpreter's face, arms, hands and fingers. They are written around a human interpreter; they do not address automated sign-language systems.

Failure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.

United States federal courts. The Court Interpreters Act says the presiding judicial officer shall utilize the services of the most available certified interpreter, or when none is reasonably available an otherwise qualified interpreter, for parties or witnesses who speak only or primarily a language other than English or have a hearing impairment, and that interpretation shall be in the simultaneous mode for any party to a proceeding instituted by the United States and in the consecutive mode for witnesses. It governs federal courts only.

Regulation, subsidy or public procurement is requiring or funding adoption — the mirror of a constraint. It shows adoption is being required, not that it has happened, so one mandate is never enough on its own; two independent ones are.

If you are starting out, aim at the settings where rules require a qualified or certified interpreter — courts, healthcare, sign language — and get the certification those settings ask for. General, low-stakes interpreting is where machines compete most.

Expect AI in your preparation and around the booth, and machine interpreting to take some low-stakes exchanges. Your certification and the rules that require it are what protect the rest.

Stay and strengthen

Stay, and certify for courts or healthcare

US federal courts must use certified interpreters where available, and US health rules require qualified interpreters.

Reshape the role

Use AI for preparation, not for the rendering

Employers such as the European Commission already guide interpreters to use approved AI tools for preparation.

Adjacent move

Move into sign language or video remote interpreting

US projections expect demand for American Sign Language interpreters to grow with video relay services.

Not where rules require a qualified person, and not yet where errors are dangerous. US federal courts must use certified interpreters, EU law requires interpretation that keeps criminal proceedings fair, and US health rules require qualified interpreters; clinical studies find machine interpreting makes far more linguistic or clinical errors. Machines do compete in lower-stakes settings and in conference research, and US projections expect interpreter and translator employment to grow 2 percent from 2025 to 2035.

We do not answer that with a number of years. Watch the rules in your setting — who may interpret in court, what health providers must offer — and whether clinical studies start finding machines as safe as people. Those will decide it more than any demo.

Under the US Section 1557 rule, a covered entity must offer a qualified interpreter when interpretation is required. In a study of three apps for simple clinical exchanges, none met the non-inferiority threshold against professional interpreters, and the authors advised clinicians to defer using them until they improve.

In regulated settings, yes: courts and healthcare require qualified people in real time, where errors cannot be revised. General commercial translation, covered on the translator page, is more exposed.