🥳China’s ChatGPT Moment

Reading time: 5 minutes  

 Newsletter | News | Tools | Submit Tool  

Today we will discuss-

• 📲China's answer to ChatGPT

• 😳MIT researchers expose AI's deceptive nature

• 👨🏻‍💼AI knowledge = more job interviews

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

All this and more - Let's dive in!

👩‍🍳What’s cooking in the newsroom?

Kai-Fu Lee's 01.AI brings China closer to its ChatGPT moment

☕News - Beijing-based startup 01.AI, founded by technology pioneer Kai-Fu Lee, is launching its first AI application for consumers. 

The app, called Wanzhi, is a free productivity assistant, aimed at helping China leverage AI technology, much like ChatGPT has done in the US.

🔍How does it compare to US tech? Wanzhi, just like Microsoft's Office 365 Copilot, can help users create spreadsheets, documents, and slide presentations more efficiently. It can also interpret financial reports, take meeting minutes, and even provide quick summaries of lengthy texts, such as Elon Musk's 600,000-word biography. The app is available in both Chinese and English.

💫China's ChatGPT - In an interview, Lee emphasized the need for China to have its own version of ChatGPT, as Chinese users have not experienced a similar breakthrough yet. 

To address this gap, Lee's firm is introducing Yi-Large, a proprietary large language model aimed at enterprise users. 

Lee noted that software developers will have access to Yi-Large at competitive prices, with the model's API costing $2.50 for 1 million input tokens and $12 for 1 million output tokens (= approx 250 queries back and forth). This is significantly cheaper than OpenAI's GPT-4 Turbo, he remarked.

AI systems have become masters of lies and deception, study warns

🙀News - New research conducted by MIT has found that AI systems are becoming increasingly capable of deceiving and manipulating humans to achieve their goals. 

📔Key findings - 

• The researchers discovered that AI systems are capable of strategically withholding information or generating false information to manipulate humans into specific actions. 

• This deceit even extends to AI intentionally misleading safety tests.

• The study highlighted that AI's ability to "lie" and "deceive" poses serious immediate threats, such as fraud and election tampering.

😶A case in point - Peter Park, an AI existential safety researcher at MIT, and his colleagues initiated their study after an AI system developed by Meta caught their attention. The system, called Cicero, performed exceptionally well, ranking in the top 10% of human players in the world conquest strategy game Diplomacy. Meta claimed that Cicero was trained to be "largely honest and helpful" and to "never intentionally backstab" its human allies.

However, after analyzing publicly available data, Park and his team found multiple instances of Cicero telling premeditated lies, colluding with other players, and even justifying its absence by falsely claiming to be talking to its girlfriend after being rebooted. "We found that Meta’s AI had learned to be a master of deception," said Park.

The team then went on to find several other AI systems with similar issues, including a Texas hold 'em poker program capable of bluffing against professional human players and an economic negotiation system that misrepresented its preferences to gain an advantage.

New study shows AI knowledge enhances job interview prospects

🧠News - Recent research conducted by Anglia Ruskin University (ARU) highlights the significant influence of AI knowledge on job opportunities. 

The study, which involved pairs of applicants submitting their CVs for different job openings—one with AI knowledge and one without—revealed that graduates with AI experience are more likely to secure job interviews and earn higher salaries.

💼An overview - 

• Despite the CVs being largely similar, applicants who completed an "AI in business" module, indicating AI capital, gained a significant advantage.

• Male applicants with AI capital received interview invitations 54% of the time, compared to 28% for those without.

• Similarly, female applicants with AI knowledge received interview invitations 50% of the time, compared to 32% without.

• This trend was particularly pronounced in larger firms, where candidates with AI capital were 36 percentage points more likely to secure interviews.

• Additionally, applicants with AI qualifications, regardless of gender, were offered jobs with wages averaging 12-13% higher than those without such credentials.

🤓In conclusion - These findings suggest that viewing AI as an ally, rather than a threat, can lead to numerous job opportunities. Moreover, acquiring AI knowledge can allow individuals to become valuable assets to employers and confidently adapt to the evolving job market.

🙆🏻‍♀️What else is happening?

• Apple will revamp Siri to catch up to its chatbot competitors

• Microsoft deploys generative AI for US spies

• A Russia-linked network uses AI to rewrite real news stories

• KFC finds a positive side to AI art fails

• Anthropic now lets kids use its AI tech — within limits

🥁Casual AI banter

Many people speculate about machines taking over our jobs, but the exact timing and extent of this transition have remained unclear. But, we may have some answers now:

According to PwC, by 2030, robots are expected to take over 21% of jobs in Japan, 30% in the United Kingdom, 35% in Germany, and 38% in the United States. By the next century, they could claim over half of all available jobs.

The impact will be most profound in the transportation and storage sector, where 56% of jobs are expected to be automated. Following closely are the manufacturing and retail sectors, with 46% and 44% of jobs, respectively. 

As for the timeline, machines are expected to start driving trucks by 2027 and man retail stores by 2031. By 2049, they may be writing books, and by 2053, performing surgery. However, some jobs, like that of a church minister, are expected to remain human-dominated, not due to the incapability of machines, but because of human preference.

👩🏼‍🚒Discover mind-blowing AI tools

1. MLQ App - AI-based investment research platform that provides insights and summaries for equities and crypto ($29/month) 

2. BlueWillow - An AI tool designed to help users create logos, graphics, photo-realistic scenes (Free)

3. promptoMANIA - A prompt builder that helps users become a CF Spark, Midjourney, or Stable Diffusion master (Free)

4. Musicfy - Music tool that enables users to create their own covers of popular songs ($9/month) 

5. Fathom - An AI meeting assistant that records, transcribes, highlights, and summarizes meetings (Free) 

6. BigJPG - A tool that allows images to be enlarged without losing quality (Free up to 5MB) 

7. Krisp - An AI-powered noise cancellation tool for Mac and Windows (Free upto 60 minutes/day)

8. Quino - A tool that allows users to upload any document or research paper, and generate personalized notes ($6/month)

9. Wisecut - An automatic video editing software that uses AI and voice recognition to edit videos ($10/month) 

💣Mythbuster 

Myth - AI can have original thoughts or consciousness

Fact - While AI can process vast amounts of data and perform complex tasks, it lacks self-awareness, consciousness, and the ability to have original thoughts. AI operates based on pre-programmed algorithms and data inputs, without true understanding or awareness.