anthropic launches openai while meta expands its platform

Key Takeaways

• Israeli brothers Meir and Yosef Nahum were indicted for using AI to create and sell false information to Iranian agents for tens of thousands of shekels.
• Anthropic's Claude chatbot was a popular topic at the HumanX AI conference, while OpenAI's ChatGPT received less attention, with some noting performance decline.
• Meta reportedly used gig workers to collect sensitive personal data, including children's photos and family conversations, for AI training, raising privacy concerns.
• Apple is developing AI smart glasses with a unique camera design, expected to launch in late 2026 or early 2027, focusing on photo and video capture.
• London's Metropolitan Police are exploring AI tools to identify child sexual abuse victims and categorize harmful imagery, aiming to improve efficiency and reduce officer exposure.
• Tulane University and Sussex Police are utilizing AI-powered cameras for surveillance, monitoring campus traffic and detecting driving offenses, respectively, which has raised privacy concerns.
• Cybersecurity experts warn of an "AI Vulnerability Storm," emphasizing that security must be integrated into AI system design from the outset to prevent overwhelming organizations with weaknesses.
• Physics Wallah partnered with Microsoft to provide AI education in Tier 2 and Tier 3 Indian cities, aiming to make AI skills accessible to a broader student population.
• Companies are investing heavily in AI for HR but often fail to achieve significant results due to issues like data quality and a lack of fundamental process redesign.

Israeli brothers indicted for selling AI data to Iran

Two Israeli brothers, Daniel and Yoni, have been indicted for allegedly selling AI-generated information to Iran. They are accused of transferring mostly false data to Iranians for money. The indictment was unsealed after a media request. The brothers reportedly used advanced methods to create and share fake information for over a year. Authorities are still investigating the full scope of the data shared.

Brothers used AI to deceive Iranian agents

Brothers Meir and Yosef Nahum were indicted for espionage after allegedly selling AI-generated information to Iranian agents. They are accused of transferring mostly false data for tens of thousands of shekels. The court allowed their names to be published, rejecting claims of harm. The indictment details how Meir used AI to create fake documents and images, including a picture of someone making an 'okay' sign, to trick the Iranian agent. He also fabricated conversations to make it seem like he had a friend in an Israeli intelligence unit.

Claude AI chatbot popular at HumanX conference

At the HumanX AI conference in San Francisco, the chatbot Claude by Anthropic was frequently mentioned. Tech professionals discussed how AI agents are changing business by automating tasks. While Claude received many mentions, ChatGPT from OpenAI was discussed less. One vendor noted a decline in ChatGPT's performance. OpenAI's CTO of B2B applications, Srinivas Narayanan, highlighted the rapid changes in AI technology, especially in software engineering.

Gig workers collected sensitive data for Meta's AI

Gig workers were reportedly asked to collect sensitive personal data, including children's photos, to train AI for Meta, Mark Zuckerberg's company. Workers on platforms like Appen and TELUS International performed tasks such as transcribing audio and labeling images. Some were asked to record family conversations or describe their own emotions. Meta stated that the data is anonymized and used responsibly. The report raises ethical concerns about privacy and the human cost of AI development.

Apple developing AI smart glasses

Apple is reportedly working on several styles of AI smart glasses with a unique camera design. This move enters a market currently led by Meta. Originally, Apple planned three head-worn devices, including AR glasses, but only the mixed-reality headset, Vision Pro, was released in 2024. Now, Apple is developing simpler smart glasses without a display, similar to Meta's Ray-Ban glasses. These are expected to launch around late 2026 or early 2027 and will focus on everyday uses like capturing photos and videos.

London police consider AI for child abuse cases

The Metropolitan Police in London is exploring the use of AI to help identify victims of online child sexual abuse and categorize harmful imagery. This technology could help find new victims faster and reduce the time officers spend viewing distressing material. Currently, officers manually review images to identify victims and categorize them by severity. The Met is discussing AI tools with tech companies to assist this process while maintaining human judgment and victim care. They are also considering technology to quickly review large volumes of messages.

Tulane University uses AI cameras, raising privacy concerns

Tulane University is using AI-powered Flock Safety cameras to monitor campus traffic by reading license plates. While intended to enhance safety, this practice has raised privacy and transparency concerns. The cameras capture vehicle data, including make, model, and color, and store it for extended periods. Experts note that these systems do more than just read plates and that advances in AI and cloud storage allow for vast data collection. Concerns exist about potential errors, misuse, and surveillance, as similar technologies have been linked to wrongful arrests and investigations.

Sussex Police use AI cameras to catch phone-using drivers

Sussex Police in the UK are using AI cameras to detect drivers using mobile phones or not wearing seatbelts. This initiative follows a successful trial in 2024 that identified hundreds of offenses. The cameras will be in place for several weeks to gather data on driver behavior and encourage change. Identified offenses will be reviewed and may result in educational courses, fines, or prosecution. Police state these cameras enhance, not replace, traditional policing efforts to improve road safety.

Leaders face AI security risks, experts warn

Cybersecurity leaders are warning of an 'AI Vulnerability Storm' due to AI accelerating the discovery of software weaknesses. Experts caution that AI-driven vulnerability discovery programs could overwhelm organizations with issues. They emphasize that security must be integrated from the start of AI system design, not treated as an afterthought. Concerns also include AI systems that learn from data, making them vulnerable to manipulation if the data source is compromised. Transparency is crucial for identifying and fixing potential weaknesses in AI systems.

Physics Wallah and Microsoft partner for AI education in India

Physics Wallah (PW), an Indian edtech company, has partnered with Microsoft to offer AI courses in Tier 2 and Tier 3 cities. This collaboration aims to make AI education more accessible and prepare students for future jobs. The courses will cover AI fundamentals, machine learning, and practical applications. Physics Wallah CEO Alakh Pandey stated their goal is to empower students from all backgrounds. Microsoft India President Anant Maheshwari highlighted the importance of AI skills for the youth.

Companies struggle to see results from AI in HR

Many companies are investing heavily in AI for HR functions like hiring and performance management, but most are not achieving significant results. Experts point to issues with data quality, governance, and a lack of process harmonization as key challenges. The main missed opportunity is how companies approach AI, often focusing on individual tasks rather than redesigning work entirely. To get meaningful results, organizations need to fundamentally rethink how humans and AI collaborate, integrating AI into the core strategy rather than just using it as a tool.

Sources

• High Court releases names of Israeli brothers indicted for selling AI-generated info. to Iran
• High Court releases names of Israeli brothers indicted for selling AI-generated info. to Iran
• At the HumanX conference, everyone was talking about Claude
• AI gig workers forced to collect personal data, including kid's pics, for Zuckerberg-backed company
• AI: Apple readies range of AI Smart Glasses. RTZ #1055
• Met considering using AI to help online child sexual abuse cases
• Surveillance or protection? AI-powered cameras at Tulane spark privacy, data security concerns
• AI cameras used to detect drivers on their phone in Sussex
• Techie Tonic: How leaders can tackle AI security vulnerabilities
• Analytics India Magazine
• HR investment in AI is booming, but most companies aren’t seeing meaningful results