Are We in a (Language) AI Bubble?

Michael Burry, the investor whose claim to fame is having called the US subprime crisis back in 2006, calls the current AI boom a bubble. Former Google CEO Eric Schmidt disagrees, saying that, if anything, AI and its implications are “underhyped.”

In between, we find current Google CEO, Sundar Pichai, who told the BBC that “while the growth of AI investment had been an ‘extraordinary moment,’ there was some ‘irrationality’ in the current AI boom.” He also said that no company would be immune to the AI bubble bursting.

We asked readers if they think AI is indeed in a bubble, and for close to half of respondents (47.8%), that is decidedly the case. For a little over a third (33.4%), it is a possibility, while the rest (18.8%) do not see an AI bubble but rather the next Industrial Revolution.

Going After the Global Classroom

Greg Hart took the helm as CEO of the publicly traded online education platform Coursera in February 2025, inheriting a company whose market capitalization had fallen nearly 80% since its 2021 IPO.

Hart’s strategy to reignite growth centers on an aggressive deployment of AI, including AI translation and AI dubbing. During the Q3 2025 shareholders call, the CEO emphasized that AI has made the “cost of translating and dubbing unbelievably cheaper,” while simultaneously improving learning outcomes. 

The company has indeed scaled quickly, growing from 100 to over 5,600 courses in 26 languages within months, and began offering AI dubbing in April 2025. The in-course GenAI tutor, “Coursera Coach,” is also now integrated into 97% of courses and available in 26 languages.

With a total learner base of 191 million, Coursera’s leadership maintains that the company is executing on its core priorities: optimizing its go-to-market strategy to capture the massive global skilling opportunity, accelerating product innovation, and expanding its content engine. To that end, it also overhauled its pricing, which Hart said was simply “out of reach” for many consumers. 

We asked readers if they think Coursera’s language AI push will work, and a little more than one in two (54.3%) think AI is only part of the solution. Over a quarter of voters (25.7%) believe it is too little too late, and one in five (20.0%) believe the strategy will work.

Hearing Arguments over the Use of AI Translation in Courts

Assembly Bill 377 (AB 377/SB 357), introduced in July 2025 before the Wisconsin State Assembly, would require official state and local government communications to be in English, with limited exceptions, and includes a provision allowing state agencies to use AI translation or “other machine-assisted translation tools” in lieu of human interpreters during court proceedings.

State Representative David Murphy asserted that integrating AI and other technologies would result in “substantial cost savings to taxpayers.” This perspective was debated during an October 22, 2025, hearing before the Committee on Science, Technology, and AI.

On the other side are those who oppose the bill, who voiced their concerns over accuracy, liability, and equitable language access for vulnerable populations. Witnesses, including students, professors, interpreters, and advocates, expressed alarm over the potential for misinterpretation in high-stakes situations.

Critics also highlighted the bill’s lack of clarity on oversight, its use of broad language (e.g., “state agencies,” “other machine-assisted translation tools”), and the ambiguity about who would actually be accountable if an AI translation error impacts a trial.

We asked readers if they think AI interpreting should be allowed in the courts, and over two-thirds of respondents (63.0%) chose unequivocally “no.” Two cohorts are equally split (15.2% each), with one group believing it depends on the case, while the other one thinks it may belong in 2030. The rest of the readers (6.6%) actually think AI should be allowed in courts.

AI-Translated and Formatted eBooks in 72 Hours?

That is Amazon’s offering with the launch of Kindle Translate. Unveiled on November 6, 2025, the AI translation service is initially being rolled out in beta to select authors in the Kindle Direct Publishing (KDP) program. 

The initial phase supports bidirectional English and Spanish, as well as German into English, with plans for expansion driven by author interest and reader demand.

The service is designed for integration within the KDP portal, allowing authors to manage the entire automated translation and publication process in one place. This includes setting prices and publishing a fully translated and formatted e-book in three days, verification steps included.

Amazon claims that a rigorous automatic evaluation process, developed in collaboration with language specialists, ensures accuracy before publication. Authors then have the flexibility to preview the resulting text.

While readers will be informed when they are engaging with an AI-translated e-book, the reception to Kindle Translate has been mixed. A few independent authors hail the service as a major breakthrough, while others express skepticism and concerns about creative control and value. Some also warn that relying on AI translation without human review poses an unacceptable risk to an author’s professional reputation.

We were curious to know what newsletter subscribers think about reading AI-translated novels, and over three-quarters (76.8%) see it as a bad idea. About one in six (16.1%) are OK with it if the author’s OK with it, and the rest (7.1%) think it is a good idea.