A recent lawsuit alleges that Perplexity AI's search engine shares user chat data, including sensitive information, with Google and Meta, even when users activate 'Incognito Mode.' The complaint claims that tracking tools embedded in Perplexity's code transmit personal identifiers and full conversation texts without user consent, potentially for targeted advertising. Perplexity has stated it has not yet been served and cannot verify these claims.
In the realm of cybersecurity, many AI security tools can classify alerts but often fall short on full investigations, lacking the ability to trace attack paths across systems. This limitation means they frequently rely on existing SIEM systems rather than fundamentally changing detection methods. The shift towards AI as a primary user of security products is making APIs more crucial than traditional user interfaces, potentially challenging the economics of expensive SIEM platforms.
AI's influence extends beyond technology, impacting workplace dynamics and governance. A new gender gap is emerging as AI shifts communication towards more explicit interactions and mediates feedback. Furthermore, many corporate boards lack sufficient AI expertise, creating significant governance and reputational risks, with few having established AI education plans or dedicated oversight committees.
Meanwhile, new initiatives aim to integrate and ethically develop AI. MIT researchers have created a framework to test the ethics of autonomous AI systems, balancing measurable outcomes with subjective values like fairness. In Japan, companies such as Lawson Inc., Sompo Japan Insurance Inc., and Itochu Corp. are incorporating AI into new hire ceremonies to familiarize employees with the technology. Carnegie Mellon University also launched a fellowship program to accelerate discoveries at the intersection of AI and astrophysics.
OpenAI is making its Codex AI model more accessible, offering flexible, pay-as-you-go pricing for Codex-only seats for teams using ChatGPT Business and Enterprise. This allows smaller groups to pilot projects and scale as needed, with usage billed based on token consumption. The annual price for standard ChatGPT Business seats has also been lowered. However, the increasing use of AI in creative fields, such as an audience member using ChatGPT to write jokes for comedian John Mulaney, raises questions about originality and human creativity.
Key Takeaways
- Perplexity AI is being sued for allegedly sharing user chat data, including sensitive information and 'Incognito Mode' conversations, with Google and Meta without user consent.
- Many current AI security tools can classify alerts but lack the capability to perform full investigations or trace attack paths across systems.
- AI is transforming security operations, making APIs more critical than user interfaces for SIEM systems and potentially challenging their traditional economics.
- A new gender gap is emerging in the workplace as AI influences communication styles, favoring explicit interactions and mediating feedback.
- Corporate boards often lack sufficient AI expertise, posing significant governance and reputational risks due to limited AI oversight and education plans.
- MIT researchers have developed a new testing framework to evaluate the ethics of autonomous AI systems, balancing measurable outcomes with subjective values like fairness.
- OpenAI has introduced flexible, pay-as-you-go pricing for Codex-only seats for ChatGPT Business and Enterprise teams, and lowered the annual price for standard Business seats.
- Japanese companies like Lawson Inc., Sompo Japan Insurance Inc., and Itochu Corp. are integrating AI into new hire ceremonies to familiarize employees with the technology.
- Carnegie Mellon University launched the Keystone Astronomy & AI Visiting Fellows Program to accelerate discoveries at the intersection of AI and astrophysics.
- Comedian John Mulaney highlighted concerns about AI's role in creative fields after an audience member used ChatGPT to write jokes for his stand-up set.
AI Security Tools Only Classify, Don't Investigate Alerts
Many AI security tools can identify and categorize alerts but cannot perform full investigations. These tools often use general AI models that lack understanding of how cyberattacks spread. To truly investigate threats, security operations need AI that can trace attack paths across different systems and analyze data from various sources. Some platforms also require users to build and maintain their own AI agents, adding complexity. A new platform called Morpheus aims to provide complete investigation capabilities, combining AI triage, attack path discovery, response orchestration, and incident tracking.
AI is Changing Security Operations and SIEMs
Attacks are happening faster, pushing security operations towards quicker responses. This trend is leading to more AI use in security, but many current AI security tools still rely on existing SIEM systems instead of changing the core detection methods. Major SIEM vendors are adapting, but the rise of AI as a primary user of security products means APIs are becoming more important than user interfaces. This shift could challenge the economics of expensive SIEM platforms. The market is still confusing, with many vendors offering similar AI solutions focused on improving analyst productivity rather than fundamentally rethinking security architecture.
Lawsuit Claims Perplexity Shares User Chats with Google and Meta
A lawsuit alleges that Perplexity's AI search engine shares user chat data, including sensitive information, with Google and Meta. This occurs even when users utilize the 'Incognito Mode,' which the suit claims is ineffective. The complaint states that opening prompts and follow-up questions are shared, and for non-subscribed users, an entire conversation can be accessed by third parties. The lawsuit claims this data sharing, including personal information, happens without user consent and violates privacy laws. Users reportedly shared financial, tax, and even health information, which could be used for targeted advertising.
Perplexity AI Sued for Sharing Chats with Google and Meta
A lawsuit filed against Perplexity AI claims its search engine sends user chat data directly to Google and Meta. The complaint states that tracking tools embedded in Perplexity's code share this information without user consent or disclosure, even when 'incognito mode' is active. The lawsuit, which also names Google and Meta as defendants, alleges that this data includes personal identifiers and the full text of user conversations. Sensitive information like financial, legal, and health details may have been shared, potentially leading to targeted advertising. Perplexity has stated they have not been served and cannot verify the claims.
AI is Creating a New Gender Gap at Work
Generative AI is changing more than just productivity and jobs; it's altering how we lead, communicate, and interact. A new gender gap is emerging as AI impacts communication styles. Cultures that rely on implicit communication are shifting towards more explicit interactions because AI requires clear prompts. AI also influences feedback, acting as a mediator that softens critiques and emphasizes positive aspects. Furthermore, AI's focus on results and examples is changing how persuasion works, moving away from traditional deductive reasoning towards more pragmatic presentations.
Corporate Boards Lack AI Expertise, Posing Reputational Risk
Many corporate boards are struggling to keep up with the rapid pace of AI-driven transformation, creating a significant governance and reputational risk. A large percentage of Fortune 100 boards have limited or no AI oversight, and many directors report having little knowledge or experience with AI. Few boards have established AI education plans or dedicated committees for AI oversight. Companies with AI-literate directors tend to outperform their peers financially. While AI can assist in board functions, human judgment remains crucial for managing risk, ethics, and reputation.
Japanese Companies Use AI in New Hire Ceremonies
Japanese companies are increasingly incorporating artificial intelligence into their new hire entrance ceremonies. Lawson Inc. used an AI mascot trained on its president's thoughts to interact with new employees. Sompo Japan Insurance Inc. created AI-generated videos showcasing the envisioned futures of its new hires. Itochu Corp. also projected AI-generated images of new employees' future selves. These initiatives aim to familiarize new employees with AI technologies and encourage them to utilize AI in their work, reflecting Japan's focus on AI development.
MIT Develops AI Ethics Testing Framework
MIT researchers have created a new testing framework to evaluate the ethics of autonomous AI systems. This system helps identify situations where AI decision-support tools might treat people or communities unfairly. It balances measurable outcomes like cost with subjective values such as fairness. The framework uses a large language model to understand stakeholder preferences and efficiently pinpoint potential ethical issues before AI systems are deployed. This approach aims to discover unforeseen problems and ensure AI aligns with human values.
OpenAI Offers Flexible Codex Pricing for Teams
OpenAI has introduced more flexible pricing for its Codex AI model, making it easier for teams to use. Teams using ChatGPT Business and Enterprise can now add Codex-only seats with pay-as-you-go pricing, avoiding fixed seat fees. This allows smaller groups to start pilot projects and scale as needed. Usage is billed based on token consumption, providing clearer cost tracking. Additionally, the annual price for standard ChatGPT Business seats has been lowered. Eligible workspaces can receive credits for new Codex team members.
Comedian Mulaney Mocks AI Joke Writing
Comedian John Mulaney shared a story about an audience member who used ChatGPT to write jokes for his stand-up set. Mulaney expressed surprise and questioned the fan's decision to rely on AI for creativity. He humorously suggested that using AI-generated jokes isn't true stand-up comedy but rather 'reading a prompt.' The anecdote highlights concerns about the increasing use of AI in creative fields and its potential impact on human originality.
CMU Launches AI and Astronomy Fellowship Program
Carnegie Mellon University has launched the Keystone Astronomy & AI Visiting Fellows Program, supported by the Simons Foundation. This initiative pairs visiting researchers with mentors in astrophysics and AI for month-long residencies. The program aims to accelerate discoveries at the intersection of AI and astrophysics by tackling high-impact problems and sharing software and workflows. It also provides opportunities for CMU graduate students to collaborate and gain experience applying AI to scientific challenges. The program will support six fellows over three years.
Sources
- The AI SOC’s L1 Automation Ceiling: Why Classification Is Not Investigation
- If AI Becomes the User, What Happens to the SIEM?
- Perplexity's "Incognito Mode" is a "sham," lawsuit says
- Bombshell AI class action suit alleges users’ chats with Perplexity go straight to Meta, Google
- AI isn’t just reshaping productivity and threatening to kill jobs. It’s changing how we lead, communicate, and treat each other. It’s also creating a new gender gap
- AI gaps in the boardroom are becoming a reputational risk
- From digital execs to future projections, Japan firms introduce AI at new-hire ceremonies
- Evaluating the ethics of autonomous systems
- Codex now offers more flexible pricing for teams
- Comedian roasts audience member's absurd usage of AI: 'What do you mean by that?'
- CMU Launches Keystone Astronomy & AI Visiting Fellows Program
Comments
Please log in to post a comment.