Here’s the current issue of The Download, our weekday digest offering a daily insight into the latest happenings in technology.
The impact of AI and Wikipedia on endangered languages
Wikipedia stands as the most extensive multilingual undertaking since the Bible: It includes versions in more than 340 languages, with another 400 lesser-known languages in progress. Yet, many of these minor editions are overwhelmed by AI-generated translations. Volunteers focusing on four African languages estimated for MIT Technology Review that around 40% to 60% of entries in their Wikipedia versions were unedited machine translations.
This is starting to create a complex issue. AI systems acquire new languages by collecting vast amounts of text from the web. Wikipedia often serves as the primary source of online linguistic information for less common languages—therefore, any inaccuracies on those pages can contaminate the resources that AI relies upon. Volunteers have been compelled to take drastic measures to rectify the problem, even eliminating certain languages from Wikipedia altogether. Discover the complete account.
—Jacob Judah
This article is a segment of our Big Story series: extensive reporting on crucial technological developments by MIT Technology Review. These narratives delve deeply into future innovations and their implications for society.Explore more from the series here.
Trump advocates leucovorin as a new autism treatment. What is it?
On Monday, President Trump asserted that childhood vaccinations and acetaminophen, the main component of Tylenol, are responsible for the rising rates of autism. He recommended that pregnant women refrain from using the medication.
The administration also revealed that the FDA would strive to make a drug called leucovorin accessible as a treatment option for children with autism. The president’s comments disappointed many observers. “The data referenced does not support the notion that Tylenol induces autism, nor that leucovorin serves as a cure—it only amplifies fear and misleadingly implies hope when there are no straightforward solutions,” stated the Coalition for Autism Researchers, a collective of 250 scientists. So what does the research indicate? Read our analysis to learn more.
—Cassandra Willyard
This is a part of our MIT Technology Review Explains series, where our writers clarify the intricate, muddled realm of technology to aid your understanding of upcoming trends. You can read more from the series here.
Fusion power facilities aren’t operational yet, yet they’re generating revenue
This week, Commonwealth Fusion Systems announced they have a new client for their inaugural commercial fusion power plant, located in Virginia. Eni, a major player in the global oil and gas sector, has entered into a billion-dollar agreement to purchase electricity from the plant.
One minor detail? That reactor is not in existence yet. This is a peculiar time in fusion energy. Investors are injecting billions into the sector to develop power plants, and companies are even committing to substantial contracts to acquire energy from these yet-to-be-built facilities.
However, all this occurs before companies have actually finalized a functioning reactor capable of generating electricity. While financial resources are essential for advancing new technology, this influx of funding may cultivate some distorted expectations. Read the complete article.
—Casey Crownhart
This feature comes from The Spark, our weekly newsletter focusing on the latest developments in climate change and clean technology. Subscribe to get it delivered to your inbox each Wednesday.
The AI Hype Index: Unpacking the chatbot phenomenon
Every day, millions of us interact with chatbots, even though we often lack an understanding of their functionality or the ways in which they influence us. In response, the FTC has recently initiated a probe into the effects of chatbots on children and adolescents. Meanwhile, OpenAI has begun to clarify what users are actually utilizing ChatGPT for, and why they believe their large language models often fabricate information.
There remains much to uncover—but that isn’t preventing governments from progressing with AI initiatives. In the US, RFK Jr. is encouraging his staff to utilize ChatGPT, while Albania is employing a chatbot for the procurement of public contracts. Explore our latest AI Hype Index edition to help distinguish AI fact from exaggerated fiction.
Essential reads
I’ve searched the web for today’s most interesting/important/scary/fascinating stories related to technology.
1 Huntington’s disease has been effectively treated for the first time
Gene therapy succeeded in slowing the disease’s progression in patients by 75%. (The Economist $)
+ Discover how the gene-editing tool CRISPR is transforming lives. (MIT Technology Review)
2 Google reports that 90% of tech employees are using AI
Yet most of them also express skepticism regarding the reliability of AI outputs. (CNN)
+ What causes AI to hallucinate? (MIT Technology Review)
3 A MAGA TikTok takeover is on the horizon
As free speech protections in the US appear increasingly fragile. (The Atlantic $)
4 Chinese tech professionals are returning from the US
Several complex factors are influencing their decision to leave, as well as their pull back. (Rest of World)
+ It remains unclear how the new $100,000 fee for H-1B visas will affect India’s tech industry. (WP $)
+ Europe is also striving to attract more tech talent. (The Verge)
5 If AI can make diagnoses, what is the role of doctors?
Medical professionals must prepare for the reality that chatbot use is expanding rapidly among patients. (New Yorker $)
+ This health startup utilizes LLMs for conducting appointments and making diagnoses.(MIT Technology Review)
6 Drones have been observed at four additional airports in Denmark
It appears to be a coordinated assault, but officials are still investigating the culprits. (FT $)
7 TSMC has revealed AI-optimized chips that are energy efficient
The AI technology discovered superior solutions compared to TSMC’s human engineers—and did so at a much quicker pace. (South China Morning Post)
+ These four visuals summarize the current state of AI and energy. (MIT Technology Review)
8 Tips for finding love on dating apps 💑
It might be challenging, but it is achievable. (The Guardian)
9 AI models struggle with Persian social customs
These customs often require saying ‘no’ when ‘yes’ is intended, which doesn’t translate well for computers. (Ars Technica)
10 VR headsets are improving, yet enthusiasm seems lacking
The tech sector frequently overestimates the willingness of consumers to wear computers on their faces. (Gizmodo)
Quote of the day
“We are experiencing the most destructive arms race in human history.”
—Ukrainian president Volodymyr Zelenskyy urges world leaders gathered at the UN to intervene and prevent the escalating development of drone technology and AI, The Guardian reports.
One last thing
The significant AI consciousness dilemma
AI consciousness transcends being a mere intellectual challenge; it poses serious moral implications. Failing to recognize a conscious AI may inadvertently subjugate an entity whose interests should be regarded. Conversely, misidentifying an unconscious AI as conscious could endanger human safety and well-being for a non-sentient, unfeeling mass of silicon and code.
In recent decades, a small group of researchers has persistently pursued understanding what consciousness entails and how it operates. This endeavor has achieved tangible advancements. Now, with the swift progression of AI technology, these discoveries may serve as our sole compass through the uncharted, ethically challenging landscape of artificial consciousness. Read the complete article.
—Grace Huckins
We can still have enjoyable experiences
A space for relaxation, enjoyment, and distraction to uplift your day. (Have any suggestions? Send me a message or let me know.)
+ It’s Fat Bear Week! Who gets your vote this year?
+ Discover Lord Woodbine, the overlooked sixth Beatle.
+ There are some truly bizarre and delightful recipes in this Medieval Cookery compilation. How about venison porridge?
+ Historical negativity towards technology is as old as technology itself, as shown by this archive.
