Connect with us

Tech News

What are large language models (LLMs), why have they become controversial?

(TECHNOLOGY) Large language models guide our AI training and recently, ethicists have pointed out serious flaws in LLMs (which cost some their jobs).



artificial intelligence large language models

“Ethical” and “AI” aren’t two words often seen together (and one of them seems rare enough on its own these days), yet artificial intelligence ethics are extremely important for all of the non-artificial beings meandering around – especially when AI has the possibility to shape and influence real-world events.

The problems presented by unethical AI actions start with large language models (LLMs) and a fairly high-profile firing in Silicon Valley.

The Morning Brew’s Hayden Field explains that large language models are machine learning processes used to make AI “smarter” – if only perceptibly. You’ve seen them in use before if you use Google Docs, Grammarly, or any number of other services contingent on relatively accurate predictive text, including AI-generated emails and copy.

This style of machine learning is the reason we have things like GPT-3 (one of the most expansive large language models available) and Google’s BERT, which is responsible for the prediction and analysis you see in Google Search. It’s a clear convenience that represents one of the more impressive discoveries in recent history.

However, Field also summarizes the problem with large language models, and it’s not one we can ignore. “Left unchallenged, these models are effectively a mirror of the internet: the good, the mundane, and the disturbing,” she writes. Remember Microsoft’s AI experiment, Tay?! Yikes.

If you’ve spent any time in the darker corners of the Internet (or even just in the YouTube comment section) you’re aware of how profoundly problematic people’s observations can be. The fact that most, if not all of those interactions are catalogued by large language models is infinitely more troubling.

GPT-3 has a database spanning much of the known (and relatively unknown) Internet; as Field mentions, “the entirety of English-language Wikipedia makes up just 0.6% of GPT-3’s training data,” making it nearly impossible to comprehend just how much information the large language model has taken in.

So when the word “Muslim” was given to GPT-3 in an exercise in which it was supposed to finish the sentence, it should come as no surprise that in over 60 percent of cases, the model returned violent or stereotypical results. The Internet has a nasty habit of holding on to old information or biases as well as ones that are evergreen, and they’re equally available to inform large language models.

Dr. Timnit Gebru, a former member of Google’s Ethical AI division, recognized these problems and teamed up with Dr. Emily Bender of University of Washington and coworker Margaret Mitchell to publish a paper detailing the true dangers of the largest language models.

Gebru and Mitchell were fired within a few months of each other shortly after the paper warning of LLM dangers was published.

There is a hilariously high number of other ethical issues regarding large language models. They take up an inordinate amount of processing power, with one model training generating up to 626,000 pounds of CO2. They also tend to grow, making that impact higher over time.

They also have a lot of trouble incorporating languages that are not specifically American English due to the majority of training taking place here, making it tough for smaller countries or cultures to develop their own machine learning at a comparable pace, which widens the gap and strengthens ill perceptions that feed into the potential for prejudicial commentary from the AI.

The future of large language models is uncertain, but with the models being unsustainable, potentially problematic, and largely inaccessible to the majority of the non-English-speaking world, it’s hard to imagine that they will continue to accelerate upward. And given what we know about them now, it’s hard to see why anyone would want them to.

Jack Lloyd has a BA in Creative Writing from Forest Grove's Pacific University; he spends his writing days using his degree to pursue semicolons, freelance writing and editing, oxford commas, and enough coffee to kill a bear. His infatuation with rain is matched only by his dry sense of humor.

Tech News

Get 8X more replies to cold emails with this affordable AI tool

(MARKETING) Cold emails are a pain in the arse – time consumption alone is a reason to let AI take over and do it for you in a personalized way!



cold emails smartwriter AI

We’ll say it: Emailing sucks, and cold emails are the epitome of misery, especially when you’re the one sending them. As unfortunate as this industry norm is, cold-emailing potential clients is a necessary evil – one that, according to SmartWriter, doesn’t have to be as arduous as one might think.

SmartWriter is a tool that automates “your entire outreach process” to take the cold email burden off of your sales team (or, if you’re a freelancer, off of yourself). This is accomplished through AI-driven writing, with leads (and the information needed to land them) being generated from email contacts, LinkedIn, and other possible links.

The thing that makes SmartWriter different from competitors, it claims, is its ability to take into account personalization attributes that are more likely to contribute to paying leads when curating its copy.

The SmartWriter process starts with selecting a type of copy to automate, with selections ranging from Instagram comments to the aforementioned cold email campaigns. Once you’ve selected a style of copy and custom options (the Instagram comment, for example, allows you to determine the type of request you want to make), you provide a sample of the writing you want to incorporate into your copy (e.g., a blog post or a LinkedIn profile link).

SmartWriter does everything else – analysis, creation of copy, implementation of custom requests and links – and then provides you with several templates. In the case of the cold email option, you can send your preferred template right from within the SmartWriter interface.

It’s a convenient response to a process that is anything but, which makes it perfect for freelancers looking to maximize their time.

SmartWriter has several different subscription options, all of which come with a free trial (with no mandatory credit card entry to boot). The cheapest option (and the one probably most effective for a small team) clocks in at just under $60 per month.

If you’re somebody who spends hours researching and curating emails to little (or no) avail, that’s a bargain price, especially when considering how much time you’ll save on the back end.

AI composition software has come a long way in the last few years, so it’s no surprise that the cold-emailing process has gone the automation route. Even if you’re one of the three people alive who doesn’t mind writing cold email templates, you still owe it to yourself (and your team) to take a closer look at SmartWriter.

Continue Reading

Tech News

Infinity Maps is the most mind-blowing visual workspace ever

(TECHNOLOGY) Infinity Maps is bringing together whiteboarding, diagramming, and real-time collaboration all in one neat tool.



infinity maps

Digital tools should be effective and efficient. They should help you plan, create, and manage your projects so your team can build solutions to your overall goals. While many tools say they are the all-in-one tool solution, this is a pretty bold statement to make. Each company is different, and one size doesn’t necessarily fit all.

However, there comes a time when such a tool comes slightly close to filling that spot. Infinity Maps seeks to do this by marrying some of the best qualities of different tools and adding its spice to the mix.

What does Infinity Maps offer?

The web application is partially an online whiteboard tool. In your workspace, called Canvas, you create your content by using cards. In these cards, you can add text, images, and files. Cards can be nested indefinitely creating hierarchies while still maintaining a “clear and concise” structure. You can do this by simply dragging a card into another card.

To visualize how each card correlates to one another, you have the option to link cards with arrows. These arrows are further organized by changing the color of each one or changing the color of the card itself.

Real-time collaboration
Infinity Maps lets your team collaborate in real-time. To work together, you can invite users to your map. When you share your workspace, you assign people different roles so they have the correct permissions to read or write on your map. Like Google’s web tools, you can see who is using the map because each username will show up next to their cursor and be assigned a different color.

Zoomable interface
Navigating through Infinity Maps is easy and works just like Google Maps. By double-clicking, you are taken directly to the card you selected. You can also scroll up and down and use the trackpad to zoom in and out of your map. This feature is super helpful when you have hundreds of cards on your map.

Why Infinity Maps?
The company says Infinity Maps is a “revolutionary new product that allows you to organize vast amounts of information visually & spatially”. It is a combination of Miro, Notion, and Google Maps all into one.

“What are we doing differently?” asks Infinity Maps CEO & Co-Founder Johannes Grenzemann. “With Infinity Maps, we are building a knowledge management system that allows you to create vast, huge knowledge bases [that] depict high complexity and depth while staying mind friendly because it’s a visual approach,” Grenzemann said.

infinity maps templates

Overall, Infinity Maps is a neat knowledge tool. It can be used in several ways, from students trying to organize their thesis to startups managing their product launches.

If you’re interested in checking them out to see if they are indeed the all-in-one tool solution, you can sign up to start mapping. A free account gives you access to 3 maps, up to 150 cards per map, and 50MB of cloud space. If you need more space to map out your ideas, you can unlock additional cards by inviting a friend or purchasing cards. Pro, unlimited, and team subscriptions plans are also available for purchase.

Continue Reading

Tech News

China cracks down on user data collection, allegedly cares about privacy

(TECH) Either China’s government just grew a conscience, or they’re trying to compete on a global stage. Either way, they’re implementing new laws.



china cares about privacy?

In an uncharacteristic looking move for end-user privacy and choice, China has passed sweeping new legislation entitled the Personal Information Protection Law. It’s set to take effect on November 1, 2021, and includes provisions governing consent in user data collection of tech applications and specifies how companies can use that data, especially if that data is to be transferred out of China.

This is the second of two pieces of legislation to emerge this year as China takes a hard look at their cyberspace and try their hand at oversight.

The Data Security law, which came into effect on Sept. 1, set classification frameworks for data based on “its economic value and relevance to China’s national security” as cited in Reuters.

According to experts, both laws will require companies to reevaluate how they collect and store data on a massive scale. As regulations continue to develop rapidly during China’s re-examination of their tech industry, companies are scrambling to meet the stringent new requirements and adjust their infrastructure for compliance at a break-neck pace.


  • The Personal Information Protection Law similar in design to Europe’s General Data Protection Regulation
  • China’s top cyberspace regulator, Cyberspace Administration of China (CAC), issued an investigation into Didi Global Inc, their version of Uber, with accusations of user privacy violations
  • An extensive set of rules targeting business practices that undermine fair competition, such as cultivating reviews, were implemented by China’s State Administration for Market Regulation (SAMR)
  • 43 apps were accused of illegally transferring user data and called out by the Ministry of Industry and Information Technology and required to make “rectifications”

Similar cyberspace scrutiny is happening in the US regarding monopolies held by some of the biggest players in tech like Google, Facebook, and Amazon but is moving very slowly through the legislative process.

In terms of how this impacts Americans, TikTok is currently one of the single most downloaded apps in the US and owned by Beijing-based company ByteDance. According to The Sun, ByteDance is now the most valuable startup in the world with an estimated value of 1 billion USD.

Many doubt that China actually cares about privacy, but some believe that keeping up the appearance of playing by modern corporate rules benefits their government as they seek global dominance.

Continue Reading

Our Great Partners

American Genius
news neatly in your inbox

Subscribe to our mailing list for news sent straight to your email inbox.

Emerging Stories

Get The American Genius
neatly in your inbox

Subscribe to get business and tech updates, breaking stories, and more!