We aim to keep our readers informed not only about events in the bitcoin industry but also about developments in related technological fields such as cybersecurity and artificial intelligence.
ForkLog AI has compiled the most significant AI news from the past week.
- Media: Tesla has agreed to launch FSD in China.
- Instagram to update its recommendation algorithm.
- Anthropic has introduced an iOS app.
- An AI avatar has become the spokesperson for Ukraine’s Ministry of Foreign Affairs.
- A gym teacher framed a school principal with a racist deepfake.
- A robot dog has been equipped with a flamethrower.
- The most important AI deals from the past week.
Media: Tesla Agrees to Launch FSD in China
Tesla shares soared by 15% following CEO Elon Musk’s visit to China. The rise was also fueled by rumors of the impending introduction of the Full Self-Driving system in the country.
Over the weekend, Tesla announced that Chinese authorities had lifted restrictions on the sale of its vehicles after the company met data security requirements. This issue had been a key obstacle to deploying FSD in China.
According to media reports, Tesla also struck a deal with Baidu, which will provide Musk’s company with access to the Chinese internet giant’s mapping and navigation technologies for use in the Full Self-Driving system.
Google Urges US Government to Revise Immigration Policy to Attract AI Specialists
Google warned the US Department of Labor that outdated immigration policies could deprive the country of access to international talent in AI and cybersecurity.
The company believes that the Schedule A list of “certified professions” lacking American workers is not flexible enough and fails to keep pace with changes. Google urged regular updates to include new professions.
“The US is one of the most challenging places to attract foreign talent, and we risk losing some of the most sought-after specialists in the world,” said a Google representative.
Companies can apply for green cards for employees, but the Department of Labor requires a PERM certificate to confirm a shortage of workers in that specialty. The process is time-consuming, so it is simplified for professions on Schedule A. Google claims that inclusion in the list speeds up document processing by about a year.
Media: Apple Actively Poaching AI Specialists from Google
Apple has lured dozens of artificial intelligence experts from Google and established a secret European lab in Zurich, Switzerland.
According to an analysis of hundreds of LinkedIn profiles, public job postings, and scientific papers, the company has been actively hiring AI specialists in recent years, particularly targeting employees of the search giant. Apple has poached at least 36 people from there since John Giannandrea took over the AI division in 2018.
Professor Luc Van Gool from ETH Zurich revealed that after acquiring two local computer vision startups, the corporation opened a research center in the city known as the Vision Lab.
Instagram to Update Recommendation Algorithm
The social network Instagram will update its recommendation algorithm to focus on original content and assist creators with smaller audiences. The change will affect aggregator accounts that copy photos and videos from other users.
Instead of reposts, the system will promote original posts if the content is relatively fresh. Authors will receive notifications when their content replaces a copy.
The ranking algorithm will also change, which previously favored popular accounts. Now Instagram will test content from lesser-known authors on small audiences, promoting successful posts more widely.
The updates are expected to help small creators gain popularity, similar to TikTok. The updated algorithms will be gradually implemented.
Anthropic Launches iOS App
AI startup Anthropic has released an iOS app providing access to the Claude chatbot. An Android version will be launched soon.
Previously, the chatbot was only accessible via the website and third-party libraries. Now users can interact with it and upload photos for analysis directly from the app.
Anthropic also introduced a new Claude Team group plan for companies with five or more people at $30 per month per seat. The plan offers more chat requests and an expanded context window.
Microsoft to Add Watermark in Windows 11 for PCs Not Meeting AI Requirements
Windows 11 will display a watermark if a computer does not meet the requirements for using artificial intelligence features. This is based on an analysis of a test build of the OS.
The warning will indicate that certain AI applications cannot be run due to an unsupported processor lacking the SSE4.2 instruction set. However, most users should not encounter this issue, as all modern chips are equipped with it.
Nevertheless, Microsoft appears to be adding this check specifically for AI applications like Advanced Copilot, AI File Explorer, and the upcoming DirectX AI Super Resolution.
AI Avatar Becomes Spokesperson for Ukraine’s Ministry of Foreign Affairs
The Ministry of Foreign Affairs of Ukraine has introduced an AI avatar “Viktoriya Shi” to deliver official comments on consular issues.
The digital twin will duplicate messages posted on the Ministry’s resources. To prevent forgery, videos will contain a QR code linking to the text version of the comment.
As stated by the Minister, Dmytro Kuleba, the use of generative AI will save the Ministry time and resources. However, the technology does not replace the official spokesperson.
The AI avatar was created by The Game Changers team with the assistance of Nazovni Tech. The digital twin is modeled after Ukrainian singer Rosalie Nombro.
AI Trained to Recognize Signs of Suicidal Thoughts from Voice
Concordia University PhD student Alaa Nfissi has developed an AI model capable of recognizing emotions in the speech of people calling suicide prevention hotlines.
Traditionally, such analysis was conducted manually by trained psychologists, requiring significant time and resources. Nfissi’s model automatically extracts speech features related to emotions by processing voice data using neural networks and recurrent blocks.
To train the AI, the researcher used real recordings of hotline calls and recordings of actors simulating various emotional states. Each audio segment was annotated according to one of four emotions: anger, neutrality, sadness, or fear/anxiety/worry.
The model demonstrated high recognition accuracy, especially for professional recordings, reaching 100% for anger. Nfissi hopes his development will help create a support system for hotline counselors.
Gym Teacher Framed School Principal with Racist Deepfake
In the US, gym teacher and former school athletic director Dajon Darien was accused of using a voice deepfake. He allegedly used the fake to frame his boss and get him fired.
The recording purportedly contains racist and anti-Semitic remarks by Pikesville High School Principal Eric Eiswert. However, independent experts determined the audio was fabricated, as it has a “consistent tone, unusually clean background, and lacks consistent breathing sounds or pauses.”
According to local media, Darien, whose name was mentioned in the clip, forged the recording. However, it is unknown what tool he used for this.
Police arrested the gym teacher at the airport. Investigators believe his motive was revenge against Eiswert, who was investigating possible misuse of school funds by Darien.
Robot Dog Equipped with Flamethrower
Company Throwflame has unveiled the Thermonator — a robot dog with a flamethrower mounted on its back. This unusual “companion” costs $9420.
According to its creators, the robot can shoot flames up to 10 meters. It is equipped with light detection, range mapping, and a laser sight. The Thermonator runs on a battery and can jump.
Manufacturers market the Thermonator as “the most powerful fire-throwing companion.” It can be used for deterring wild animals, agricultural work, environmental protection, snow and ice removal, as well as for entertainment and special effects.
Most Important AI Deals of the Past Week
From April 29 to May 3, AI startups raised over $1.3 billion. Here are the most interesting deals.
- Tether acquired a controlling stake in brain-computer interface manufacturer Blackrock Neurotech for $200 million.
- CoreWeave, which creates GPU-based cloud computing, raised $1.1 billion.
- Startup Lamini, which develops a platform to help enterprises deploy generative AI technologies, secured $25 million.
- Medical AI startup AISAP received $13 million in seed funding.
Also on Forklog
- Report: Tesla Autopilot and Full Self-Driving linked to hundreds of accidents and dozens of fatalities.
- Speculation arises over GPT2-chatbot as a potential successor to GPT-4 from OpenAI.
- AMD aims to sell $4 billion worth of AI chips by the end of the year.
- American newspapers have sued OpenAI and Microsoft over copyright violations in AI training.
- Vienna conference calls for regulation of AI weapons.
- Elliptic has trained AI to detect money laundering through bitcoin.
- Microsoft invested in OpenAI out of fear of falling behind Google.
- Sam Altman: Useful AI agents will become a killer feature.
- Sora from OpenAI generated a music video for an indie musician.