Might AI translation inadvertently shrink reality itself?

Google Gemini CC0

AI language translation is making communication easier around the world. But might it also shrink our very perception of reality?

That's the question I asked of a thought-provoking panel on language acquisition at the 17th Global Legal Skills Conference at the Faculty of Law at Masaryk University (MUNI) in Brno, Czechia, Thursday.

Panelists discussed the question whether foreign language acquisition will remain a virtue in a world in which artificial intelligence increasingly makes communication seamless across borders.

Karen Lundquist, University of Minnesota Law, answered aptly, in sum, "It depends." Certainly there are client interactions she can now have with AI translation tools, Lundquist said, that might not have been possible before. But those conversations do not perfectly replicate connection in a shared language. At least for now, MUNI linguist Kateřina Chudová said, there are non-verbal or near-verbal properties of communication that even AI cannot bridge, such as body language, cultural context, and irony.

Will AI get there? Probably, as fast as the technology is evolving. Attorney Luca Forgione observed that the world today looks ever more like the fictional Minority Report, a film that came out more than 20 years ago (2002). (I noted in the Q&A that Philip K. Dick published the story "The Minority Report" in Fantastic Universe in 1956!)

Disclaiming that I am no expert in linguistics, I asked in the Q&A that the panelists might reflect on the science showing that language and a person's very perception of reality are causally interactive, that is, something of a chicken-or-egg problem. For example, I said, it's almost self-evident that how a language uses tense is indicative of how a culture understands time.

In a world of AI translation, then, are we on the cusp of a global cultural convergence? More to the point: Will AI universalism cost humanity a multiplicity of realities? 

My answer to the central question on the panel, whether foreign language acquisition still has value, is an emphatic yes, for much the reasons the panelists posited. I'm no polyglot, but with just one other language and a smattering beyond, I understand the powerful link between speech and thought. Thinking about how I would conjugate a verb in Spanish sometimes helps me to rethink how best to say it in English. For example, the imperfect tense in Spanish evokes a sense of time with no precise English equivalent.

Forgione said in response that even not knowing all of the six languages his wife and children know, he can detect differences in tone that correlate to language, especially in emotionally laden contexts. Having various languages in my extended family, I understand that. Lundquist, who lived many years in Rome, suggested that Italian, for example, possesses a richer capacity than some other languages to communicate emotional intensity. That's a controverted proposition. Yet it does feel credible to me, remembering my Italian relatives and some of the language.

The differences might run deeper than the merely interpersonal. The question I asked the panel in Brno about the science of language was informed by a memorably haunting episode of Radiolab from more than a decade ago. "Colors" (2012) is widely regarded as a classic installment of the groundbreaking podcast.

To put criminally concise description to but one proposition of "Colors": Analysis of poetry suggests that ancient peoples might not have perceived the color "blue" before they acquired a word for it. With no concept of "blue," they didn't describe the sky that way. But it's not just a problem of description. With no human concept of blue, did blue even exist?

I know, you're thinking, well, there was still light at the shorter-wavelength end of the visible spectrum. But dig a little deeper. It's really a variation on the tree-falls-in-the-forest problem. And the answer is important, because in a world of intelligent—dare I say sentient?—machines, it's becoming less clear whose perception gets to define reality.

After the panel, I was fortunate to meet Lindsey Kurtz, a linguist and teacher at Penn State Dickinson Law. She taught me that the scientific concept I was after is called "the Sapir-Whorf Hypothesis," or "linguistic relativity." Learn more at the History and Philosophy of the Language Sciences podcast (2023) or at the Lingthusiasm podcast (2025).

The idea, restated by anthropologist Edward Sapir in 1929, is that structures of language are interrelated with perception and thought themselves. There are weaker and stronger versions of the hypothesis to describe the depth and inextricability of the interrelationship. 

The implications quickly become surreal. Does something exist before the mind can describe or memorialize it? The answer might be no. That is, it's possible that words are not a consequence of reality. At least sometimes, words might cause reality, that is, bring it into existence. Language is literally creative.

So what happens if language differences go away? What if artificial intelligence causes a convergence of the human diversity expressed in language, leaving behind only monochromatic modules of machine-readable meaning for our consumption? 

Is it possible, then, that humans will lose the ability to create new concepts? that our creative well will run dry? Or worse, might we inadvertently and irrevocably transfer our creative power to AI? Will AI create new planes of reality beyond our comprehension and leave us behind to wallow in the blissful ignorance of "the matrix" (the "simulation hypothesis")? Are we in it already?

In Brno, or at least in the Brno matrix, there was nary an objection to the proposition that foreign language acquisition continues to have value for learners, including lawyers, professors, and law students. As yet, there is no perfect proxy for language to effect a meeting of human minds.

Yet Star Trek's "universal translator," or Doctor Who's "voice integrator," is every day less a fiction. And that appealing Utopian imagining might camouflage a grim threat to the infinitude of humanity.

The Global Legal Skills Conference is a project of the Global Legal Skills Institute. Conference and institute are passion projects of a long-time colleague I greatly admire, Mark Wojcik, Illinois Law, whom I first met in the Association of Legal Writing Directors (ALWD) in 1997; and others in his coterie, including the sharp-minded Lurene Contento, Chicago-Kent Law, who moderated the panel: "Is Language Acquisition Still a Valuable Global Legal Skill?"

Colleagues and I presented on another panel at the conference, which I wrote about on June 2.

Global collab promotes teaching law without borders

Peltz-Steele, Lewinbuk, Rott-Pietrzyk, Kim, Rigó
© Used with permission

Collaborators and I had the privilege of discussing the Global Law Classroom (GLC) at the 17th Global Legal Skills Conference (GLS 17), hosted by the Global Legal Skills Institute and the Faculty of Law at Masaryk University (MUNI Law) in Brno, Czechia, last week.

The GLC is a collaboration of faculty around the world to bring together students across borders, via Zoom, to study international and comparative law and learn from each other. I wrote about the GLC here at The Savory Tort about a year ago, after a 2022 pilot run and just before we executed the first official program over eight weeks in fall 2024. I used the GLC as a one-credit component of my Comparative Law class, and I will again in the upcoming fall 2025.

© Used with permission

© Used with permission

In presenting on the GLC to our GLS 17 colleagues in Brno on Thursday, we provided a demonstration hypothetical in data protection for attendees to discuss in small groups. I developed the fact pattern initially with Cristina Blasi Casagran, Autonomous University of Barcelona, and we used it in the fall 2024 GLC human rights module to demonstrate divergence in U.S. and EU approaches to privacy.

For GLS 17, I created a video narrative (below) and briefed the audience on the salient doctrine of the respective legal systems (inset below video) (both CC BY-NC-SA 4.0).

RJ Peltz-Steele CC BY-NC-SA 4.0

The GLC is the brainchild of Melanie Reid, Lincoln Memorial Law, who could not join us in Brno, but will lead a discussion of the project later this summer at the annual conference of the Southeastern Association of Law Schools (SEALS).

The GLC faculty team in Brno was led by Rosa Kim, Suffolk Law, and also comprised Katerina Lewinbuk, South Texas College of Law; Balázs Rigó, Eötvös Loránd University (ELTE) (Hungary), and Ewa Rott-Pietrzyk, University of Warsaw (Poland). It was great fun for the five us to be together IRL after so much labor together on Zoom.

I'll have a another report from GLS 17 here at The Savory Tort on Wednesday, June 4.

In parting meditation on pub gossip, Czech judge peels onion on privacy limits, judicial transparency

Does GDPR pertain to pub buzz?, AG Bobek asks.

(Skeptics in the Pub Oslo (2015) by Dizzi90 CC BY-SA 4.0)

Earlier this month, Czech judge and legal scholar Michal Bobek rounded out a six-year term as an Advocate General (AG) of the European Court of Justice with a mind-bending meditation on the ultimate futility of enforcing data protection law as written and a confirmation of the essentiality of transparency in the courts.

The case on which Bobek opined hardly required a deep dive.  He said so: "This case is like an onion," he wrote.  "I believe that it would be possible, and in the context of the present case entirely justified, to remain at that outer layer.   No peeling of onions unless expressly asked for."

But the case provided Bobek an optimal diving board, and, on the penultimate day of his term as AG, he plunged and peeled.

Complainants in the case were litigants before the Dutch Council of State (Raad van State).  They asserted that disclosure to a journalist of summary case information, from which they could be identified and details of their personal lives worked out, violated their right of privacy under the General Data Protection Regulation (GDPR) of the European Union, as transposed into Dutch law.

The disclosures are permissible under a GDPR exemption for judicial activities, Bobek concluded.  But en route to that conclusion, he further opined that the potentially unbridled scope of the GDPR must be tamed to accord with social norms and democratic imperatives.

With remarkably plain reasoning, he framed the problem in a comfortable venue:

If I go to a pub one evening, and I share with four of my friends around the table in a public place (thus unlikely to satisfy the private or household activity exception of ... the GDPR) a rather unflattering remark about my neighbour that contains his personal data, which I just received by email (thus by automated means and/or is part of my filing system), do I become the controller of those data, and do all the (rather heavy) obligations of the GDPR suddenly become applicable to me? Since my neighbour never provided consent to that processing (disclosure by transmission), and since gossip is unlikely ever to feature amongst the legitimate grounds listed in ... the GDPR, I am bound to breach a number of provisions of the GDPR by that disclosure, including most rights of the data subject contained in Chapter III.

The pub might not be the only place where the GDPR runs up against a rule of reason.  Consider the more nuanced problem of footballers considering a challenge against the processing of their performance stats.  Goodness; the pub convo will turn inevitably to football.

Let's step back for a second and take stock of the GDPR from the perspective of the American street.

Americans don't get many wins anymore.  We just retreated from a chaotic Afghanistan, despite our fabulously expensive military.  We resist socialized healthcare, but we make cancer patients finance their treatments on Go Fund Me.  We force families into lifelong debt to pay for education, undermining the social mobility it's supposed to provide.  We afford workers zero vacation days and look the other way from the exploitation of gig labor.  Our men's soccer team failed to qualify for the last World Cup and Olympics, while we're not sure why our women are rock stars; it can't be because we pay them fairly.  When it comes to personal privacy, we tend to want it, but our elected representatives seem eager to cede it to our corporate overlords.

Truth be confessed, then, Americans are willing to engage in a smidge of schadenfreude when Europeans—with their peace, their healthcare, their cheap college, their Ryanair Mediterranean vacations, their world-class football, and their g—d— G—D—P—R—get themselves tied up in regulatory knots over something like the sufficient size of a banana.  Ha.  Ha.

Therein lies the appeal, to me, of Judge Bobek's train of thought.  He finds inevitable the conclusion that posting case information is data processing within the purview of the GDPR.  The parties did not even dispute that.  For today, Bobek found an out through the GDPR exemption for the business of the courts in their "judicial capacity."

The out required a stretch to accommodate posting information for journalists, which is not, most strictly speaking, a judicial capacity.  Bobek reasoned by syllogism:  For the courts to do what they do, to act in the judicial capacity, they require judicial independence.  Judicial independence is maintained by ensuring public confidence in the judiciary.  Public confidence in the judiciary is bolstered by transparency in the courts.  Transparency in the courts is facilitated by the provision of case information to journalists.  Therefore, the judicial capacity requires publication of case information to journalists.

The problem, tomorrow, is that there is no answer in the case of pub gossip.  Bobek meditated on the human condition: "Humans are social creatures.  Most of our interactions involve the sharing of some sort of information, often at times with other humans. Should any and virtually every exchange of such information be subject to the GDPR?"

Bobek

Can't be, he concluded.

[I]n my view, I suspect that either the Court, or for that matter the EU legislature, might be obliged to revisit the scope of the GDPR one day. The current approach is gradually transforming the GDPR into one of the most de facto disregarded legislative frameworks under EU law. That state of affairs is not necessarily intentional. It is rather the natural by-product of the GDPR's application overreach, which in turn leads to a number of individuals being simply in blissful ignorance of the fact that their activities are also subject to the GDPR. While it might certainly be possible that such protection of personal data is still able to "serve mankind," I am quite confident that being ignored as a result of being unreasonable does not in fact serve well or even contribute to the authority or legitimacy of any law, including the GDPR.

While we await reassessment of the bounds of data privacy law in modern society, Bobek opined more and mightily on the importance of judicial transparency as a countervailing norm.  He opened the opinion with philosopher-jurist Jeremy Bentham:

"Publicity is the very soul of justice. It is the keenest spur to exertion, and the surest of all guards against impropriety.… It is through publicity alone that justice becomes the mother of security. By publicity, the temple of justice is converted into a school of the first order, where the most important branches of morality are enforced...."

Bobek later picked up the theme:

Judging means individualised detail brought to the public forum....

On the one hand, the basis for judicial legitimacy in an individual case are its facts and details. The judge settles an individual case. His or her job is not to draft abstract, general, and anonymous rules detached from individual facts and situations. That is the job of a legislature. The more a judicial decision departs from or hides the factual background to a public court case, or if it is later reported with significant limitations, the more often it becomes incomprehensible, and the less legitimate it becomes as a judicial decision.

On the other hand, ever since the Roman age, but presumably already earlier, if a claimant asked for the help of the community or later the State to have his claim upheld and enforced by the State, he was obliged to step into the public forum and let his case be heard there. In classical Roman times, the applicant was even entitled to use violence against the respondent who refused to appear in the public (the North Eastern part of the Roman Forum called comitium), before the magistrate (seated on a rolling chair on a tribune higher than the general public—hence indeed tribunal), when called before a court (in ius vocatione).

It is true that, later on, there were other visions of the proper administration of justice and its publicity. They are perhaps best captured by a quote from a judge in the Parlement de Paris writing in 1336 instructions to his junior colleagues, and explaining why they should never disclose either the facts found or the grounds for their decision: "For it is not good that anyone be able to judge concerning the contents of a decree or say 'it is similar or not'; but garrulous strangers should be left in the dark and their mouths closed, so that prejudice should not be caused to others.... For no one should know the secrets of the highest court, which has no superior except God...."

In the modern age, returning to the opening quote of Jeremy Bentham, it is again believed that even garrulous strangers should be allowed to see and understand justice. Certainly, with the arrival of modern technologies, a number of issues must continuously be re-evaluated so that garrulous strangers cannot cause prejudice to others....

Naturally, the publicity of justice is not absolute. There are well-grounded and necessary exceptions. The simple point to keep in mind here is: what is the rule and what is the exception. Publicity and openness must remain the rule, to which naturally exceptions are possible and sometimes necessary. However, unless the GDPR were to be understood as imposing a revival of the best practices of the Parlement de Paris of the 14th century, or other elements of the Ancien Régime or the Star Chamber(s) for that matter, it is rather difficult to explain why, in the name of the protection of personal data, that relationship must now be reversed: secrecy and anonymity were to become the rule, to which openness could perhaps occasionally become the welcome exception.

Bobek seems content with judicial exceptionalism in the GDPR framework.  I'm not so sure.  I rather think the problem of the courts points to the broader problem of GDPR scope.  Will there ultimately be a pub exception, too?  Stubborn American insistence on framing data protection as business regulation, as in California data protection law, suddenly exhibits some appeal.

The case is X v. Autoriteit Persoonsgegevens, No. C-245/20, Opinion of Advocate General Bobek (Oct. 6, 2021).  HT @ Edward Machin, writing in London for Ropes & Gray.

This is not Bobek's first high-profile opinion on the GDPR—even this year.  Read in Fortune about his January opinion in a Facebook case.