Sponsored by

Reading time: 5 minutes  

 Newsletter | News | Tools | Grow My Tool  

AI-Powered Creativity

Instant, polished presentations powered by AI. Impress your audience effortlessly with Gamma. Engage users on any device. Measure engagement, get quick reactions, and collaborate seamlessly.

Try for free.

Today we will discuss-

• 🔫U.S. army testing AI chatbots in war games

• 🧯AI companies under fire for hindering independent safety testing

• 🩺AI chatbots may improve depression treatment

• ⚙️9 amazing AI tools you might not have heard of

All this and more - Let's dive in!

👩‍🍳What’s cooking in the newsroom?

U.S. Army tests AI chatbots as battle planners in a war game simulation

🎮News - The U.S. Army Research Laboratory is exploring if OpenAI's technology can improve battle planning, although only in the context of a military video game.

🔬The experiment - According to a recent report, U.S. Army researchers are using OpenAI's GPT-4 Turbo, GPT-4 Vision and two other  models with older technology to gather information about simulated battlefield terrain, friendly and enemy forces, and military tactics. Following this, they assigned the AI assistants a mission to eliminate all enemy forces and capture a specific objective point.

The study found that while OpenAI's GPT models performed better than the other two, they also led to more casualties while carrying out mission objectives. 

⭐What is the most important point here? Generative AI is just a small aspect of the U.S. Army's broader initiative to fully embrace AI in their strategies. Last week, OpenTools reported that the military had confirmed using AI for analyzing battlefield images and even identifying targets for airstrikes under Project Maven. 

And even though relying on machines to make life-or-death decisions sounds like something out of a Terminator movie rather than a promising AI future, the military is moving forward. The Pentagon has reportedly asked for billions of dollars from U.S. lawmakers to advance its AI and networking capabilities. Additionally, they've created new roles like Chief Digital and AI Officer to help implement and spread this tech across the department.

Top AI Researchers slam OpenAI, Meta, and others for hindering independent safety testing 

👨🏻‍🏫News - Over 100 leading AI researchers have penned an open letter urging companies developing generative AI to grant investigators access to their systems. They contend that restrictive company policies hinder safety testing of tools used by millions of consumers.

📝What does the letter say?

• The researchers believe that overly strict protocols intended to prevent AI system misuse are discouraging independent research. [Here it is important to note that to conduct independent investigations, researchers often have to break these protocols. And because they use their own accounts, there is a concern that AI companies, which are still devising methods to monitor rule breakers, might target users who draw negative attention to their business.]

• The letter, addressed to companies like OpenAI, Meta, Anthropic, Google, and Midjourney, urges tech firms to establish a legal and technical safe harbor for researchers to investigate their products. 

• It emphasizes that generative AI companies should learn from the errors of social media platforms, which have restricted certain types of research aimed at holding them accountable.

• Despite companies like OpenAI offering special programs for researcher access, the letter contends that this approach promotes favoritism, as companies hand-pick their evaluators. It suggests that then the only effective means to influence change is through "gotcha" moments when the company is caught making mistakes.

AI chatbots outperform doctors in following depression treatment guidelines

👨‍🔬News - According to a study conducted by a team of researchers in Israel, AI chatbots such as ChatGPT might be better at following depression treatment standards than human doctors.

The study published in the journal Family Medicine and Community Health, compared the responses of ChatGPT versions 3.5 and 4 to those of 1,249 French primary care doctors. It evaluated their responses to hypothetical patients with varying characteristics like gender, socioeconomic status, and symptoms, including mild to moderate depression.

🤓What were the findings?

• In cases of mild depression, ChatGPT-3.5 and ChatGPT-4 advised psychotherapy in 95% and 97.5% of cases, while primary care doctors recommended psychotherapy in only 4.3% of cases, prescribing drugs 48% of the time and a combination of psychotherapy and medication in 32.5% of cases.

• In severe depression cases, 44.5% of doctors suggested a combination of psychotherapy and drugs, while ChatGPT 3.5 and 4 recommended this in 72% and 100% of cases, respectively.

• Regarding medication recommendations, ChatGPT-4 favored exclusive antidepressants in 74% of cases, and ChatGPT-3.5 in 68%, while human doctors preferred a combination of antidepressants and anxiolytics/hypnotics in 67.4% of cases.

😬So what's the final verdict? According to the researchers, chatbots have the potential to improve primary care decisions as they were seen to be free from gender or social class biases that are often present in human decision-making. However, additional research is necessary to improve AI recommendations for severe cases and address potential risks and ethical concerns that may arise from their use in practice.

🙆🏻‍♀️What else is happening?

• Competition in AI video generation heats up as DeepMind alums unveil Haiper

• OpenAI and Elon Musk

• Google is starting to squash more spam and AI in search results

• China to step up quantum computing, AI in tech self-sufficiency drive

• Perplexity poised to become latest AI startup to hit Unicorn status

😎Cool AI inventions

• AI-powered chess board - The Square Off Grand Kingdom Chess set takes chess to a new level with cutting-edge robotics and Artificial Intelligence for automated moves, enhancing the gaming experience. Its companion app allows players to compete against human opponents or challenge the intelligent board. With a range of 20 difficulty levels, it accommodates players of varying skill levels.

• Implantable payment chips - In 2021, Walletmor, a British-Polish company, became the first to offer implantable payment chips. These chips enable hands-free payments by allowing users to make transactions simply by placing their hand near a contactless card reader, eliminating the need for wallets, cards, or phones. This technology is also expected to extend into applications such as house keys and car keys.

• AI-powered shoes - Moonwalker, developed by Shift, are AI-powered shoes that allow you to walk at running speed, i.e., 2.5 times faster than usual, with no extra physical exertion. These shoes distinguish themselves from skates with a lock mode for scenarios such as stairs, buses, trains, or any situation where a running pace is not desired.

👩🏼‍🚒Discover mind-blowing AI tools

1. goHeather - An AI-powered contract review software that helps users understand the terms and potential risks in their contracts ($29.99/month)

2. Larry the Elf - A tool that assists users in finding gifts for various recipients, such as children, friends, or spouses (Free)

3. Ankara AI - An app that uses AI to generate narrations for videos (Free)

4. Medisearch - An AI-powered tool that provides quick and accurate answers to health and bioscience questions (Free)

5. GPTinf - A tool designed to rephrase AI-generated content to bypass detection by AI plagiarism detectors ($12/month)

6. UpCodes - A tool that helps professionals in the construction industry stay up-to-date with building codes and regulations ($39/month)

7. Pseudoface - Generates artificial faces that can be used by content creators who want to remain anonymous (0.99/month) 

8. Face26 - Helps to improve, enhance and fix the quality of your photos in just one click ($72/year) 

9. Supermeme - An AI-powered meme generator that helps users create memes from any text ($9.9/ month) 

🕰️Historic moments 

• 1993 - Author and computer scientist Vernor Vinge published an article that popularized the idea of singularity. It is considered a point at which machines become smarter than humans. Well, this hasn’t happened yet.

• 1972 - The first intelligent humanoid robot was built in Japan which was named as WABOT-1.

• 2003 - AI scientist Yoshua Bengio and his team developed the first feed-forward neural network language model. This predicted the next word when given a sequence of words.