AI in Marketing

The emergence of agentic AI, marked by tools like OpenClaw, Antigravity, and Claude, is reshaping industries and raising concerns about job security and ethical use. OpenClaw, an open-source tool, allows deep system access for tasks like inbox management and content curation, akin to a robotic assistant with autonomy over user data. Antigravity, developed by Google, functions as a coding agent, streamlining the development process by acting like a junior developer capable of building and testing applications. Claude, from Anthropic, focuses on automating legal tasks, impacting legal-tech and SaaS markets by offering specialized domain knowledge in legal and financial sectors. The potential for misuse of these AI agents is significant, as they require substantial trust in providers to prevent data leaks or unfair advantages. Open-source tools like OpenClaw lack centralized oversight, complicating governance. Despite these risks, agentic AI presents opportunities to reduce human cognitive load by handling routine tasks, allowing humans to focus on more complex activities. Ensuring responsible AI use involves implementing accountability, transparency, and security measures, along with a shared ontology to standardize communication across diverse systems. By establishing ethical guidelines and a trust framework, these agents can become valuable assets in enhancing productivity and efficiency across various domains.

Next.

Hackers have leaked the source code for Anthropic's Claude Code on GitHub, embedding malware in some copies. Anthropic is issuing takedown notices to remove these copies. The FBI has classified a cyber intrusion into its surveillance systems as a major incident, suspecting Chinese involvement. This breach, involving unclassified systems with sensitive data, highlights ongoing vulnerabilities in the FBI's infrastructure. A 22-year-old student, Benjamin Brundage, played a key role in dismantling four major botnets, including Kimwolf, by gathering intelligence and collaborating with law enforcement. Meanwhile, North Korean hackers are suspected of stealing $280 million from the Drift crypto platform, contributing to nearly $300 million in crypto thefts this year. Cisco has also fallen victim to a supply chain attack by the TeamPCP hacker group, which compromised security software to steal source code and user credentials. In other news, Apple released patches for iOS 18 to combat the DarkSword hacking technique, which infects iPhones via malicious websites. The US-Israel conflict with Iran continues, with Iran threatening US tech companies. Shipping disruptions in the Strait of Hormuz are impacting the global economy, and concerns are rising about potential damage to Iran's nuclear facilities from US strikes.

Meanwhile.

Rivian is set to introduce its AI assistant to its electric vehicles in early 2026, following a two-year development effort. This assistant will be available across all existing models, including the R1T truck and R1S SUV, and will allow users to control climate settings and other infotainment tasks. It will also integrate with third-party apps, starting with Google Calendar, through an agentic framework developed by Rivian engineers. The AI assistant will utilize advanced large language models like Google Vertex AI and Gemini for natural conversation and reasoning. Rivian CEO RJ Scaringe's strategy to enhance vertical integration was highlighted at the company's AI & Autonomy event in Palo Alto, California, where the company also showcased new software and hardware developments, including a custom 5nm processor created with Arm and TSMC. This processor will enhance Rivian's hands-free driving assistance system, potentially allowing drivers to divert their attention from the road. Rivian's software team, led by Wassym Bensaid, has been working on a model- and platform-agnostic architecture branded as Rivian Unified Intelligence (RUI). This architecture includes custom models and an orchestration layer to ensure seamless operation of various AI models. RUI will not only power the AI assistant but also improve vehicle diagnostics, acting as an expert assistant for technicians by analyzing telemetry and history to identify complex issues.

On a different note.

AI-generated videos have become a tool in Hungary's election campaign, with Prime Minister Viktor Orbán's Fidesz party using them to target rival Péter Magyar. A controversial video falsely depicting a Hungarian soldier's execution was shared by Fidesz, alleging Magyar would involve Hungary in the Ukraine war. Despite these tactics, Magyar's Tisza party leads in polls. Fidesz's campaign includes claims that Magyar supports Ukraine financially and would impose conscription, which Tisza denies. AI videos have also been used to depict European Commission President Ursula von der Leyen in discussions with Magyar. Relations between Hungary and Ukraine have soured, with anti-Ukrainian sentiment rising. Magyar counters Fidesz's narrative through social media, appealing to younger voters.

In other news.

OpenAI is undergoing significant leadership changes as Fidji Simo, CEO of AGI deployment, takes medical leave for several weeks. Greg Brockman, OpenAI's president, will oversee product teams during her absence. Simo, who joined OpenAI in August 2025, previously managed consumer-facing products like ChatGPT and Codex. She recently closed the Sora app to focus on core products, aligning with OpenAI's potential IPO plans. Brad Lightcap, the chief operating officer, is moving to a special projects role, overseeing forward-deployed engineers. Kate Rouch, the chief marketing officer, is also on medical leave due to breast cancer treatment and will return in a more focused role. OpenAI is searching for a new CMO and a chief communications officer following Hannah Wong's departure in January. Chris Lehane is temporarily leading the communications team. An OpenAI spokesperson stated the company remains committed to advancing research, expanding its global user base, and enhancing enterprise applications. OpenAI recently raised $122 billion, valuing the company at $852 billion, marking the largest funding round in tech history.

Shifting gears.

OpenAI is expanding its ChatGPT Ads by launching self-serve access, prompting PPC managers to evaluate its potential as a new acquisition channel. Initially, the ads were part of a limited, invite-only pilot in the U.S., which was expensive and not widely adopted. However, OpenAI's move to self-serve capabilities and expansion into countries like Canada, Australia, and New Zealand signals a shift towards broader accessibility. Despite generating over $100 million in annualized revenue from the pilot, the ads' effectiveness remains uncertain, with low click-through rates compared to other platforms. Advertisers are interested, but the platform's long-term value as a marketing channel is still unproven. The self-serve model may change this, allowing more advertisers to test its viability. ChatGPT Ads could become significant if they align with user behaviors in research and decision-making, but the current hype may outpace the platform's demonstrated capabilities.

After that.

The Pentagon has formalized Palantir's Maven AI as a core military system, securing multi-year funding and increasing its investment to $13 billion from $480 million in 2024. Maven, an AI-enabled targeting platform, is now designated as an official program of record, ensuring stable funding for its development and integration across U.S. combatant commands. This designation places Maven in the Future Years Defense Program, providing budget stability that experimental programs lack. The U.S. Army will manage Maven contracts, with oversight transferring from the National Geospatial-Intelligence Agency to the Chief Digital and AI Officer within 30 days. Maven began in 2017 to apply machine learning to drone surveillance, initially partnering with Google, which withdrew in 2018 after employee protests.

Finally.

Neuralink's Blindsight, announced by Elon Musk in 2024, is a brain-implant device designed to restore vision in blind individuals by transmitting visual information directly to the brain. Unlike traditional retinal implants, Blindsight is implanted in the visual cortex, where it processes visual data through neuron stimulation using patterns from an external camera. Initial tests on monkeys have shown promising results, with one monkey having the implant for three years. Despite being designated a "breakthrough device" by the FDA, Blindsight is not yet deemed safe or effective, and must undergo clinical trials before full approval. The technology faces challenges, including the longevity of the implant and the risks associated with brain surgery. Skepticism persists due to limited information and the absence of details on Neuralink's website. Musk acknowledged that initial vision quality would be low, akin to Atari graphics, but suggested future enhancements could surpass natural vision, potentially allowing perception in infrared and ultraviolet. Experts caution that visual prosthetics are assistive rather than curative, and restoring natural vision remains a significant challenge. Philip Troyk, a biomedical engineer, notes the potential of Neuralink's technology but criticizes the messaging surrounding its capabilities.