AI in Marketing

Anthropic has developed a new artificial intelligence model, Claude Mythos Preview, which it claims is too powerful to release publicly due to its potential cybersecurity implications. Instead, the company is collaborating with over 40 technology firms, including Apple, Amazon, and Microsoft, to utilize the model for identifying and patching security vulnerabilities in critical software. This initiative, named Project Glasswing, aims to preemptively address potential threats posed by advanced AI capabilities. Anthropic's chief science officer, Jared Kaplan, emphasized the importance of raising awareness and equipping "good actors" to secure both open-source and private infrastructure. The model has already identified numerous bugs and vulnerabilities, including a 27-year-old flaw in OpenBSD and issues in widely used video software. Anthropic's decision to withhold the model from public release echoes a similar move by OpenAI in 2019 with its GPT-2 model. The company, which has seen its annual revenue triple to over $30 billion, is positioning Claude Mythos Preview as a pivotal tool in cybersecurity, capable of autonomous security research and identifying zero-day vulnerabilities. Cybersecurity experts, including Elia Zaitsev from CrowdStrike, acknowledge the model's potential as both a defensive asset and a risk if exploited by adversaries. The ongoing development of such AI models is expected to intensify the cybersecurity arms race.

Next.

The AI Alliance has launched Project Tapestry, an open-source platform aimed at fostering global collaboration in AI development while maintaining local control and sovereignty. The initiative seeks to enable distributed training of advanced AI models, allowing institutions, industries, and nations to contribute to and build upon a shared foundation without relinquishing data control. Yann LeCun, a Turing Award laureate and prominent figure in AI, joins as Chief Science Advisor to guide the project's scientific direction. Project Tapestry addresses the concentration of AI development within a few companies and regions by offering a collaborative approach that allows participants to create sovereign derivative models tailored to their specific needs and governance frameworks. The AI Alliance, a nonprofit organization with over 200 member organizations, will support and host the Tapestry platform, aiming to develop an open global model that leverages broader expertise and resources. The project is expected to overcome barriers in compute, funding, and talent that many sovereign AI efforts face. A workshop in Paris on May 7-8 will gather global technical leaders to define the project's architecture and priorities. The AI Alliance emphasizes that open innovation can lead to more capable and accountable AI, with Project Tapestry representing a significant step toward collaborative AI infrastructure development.

On that note.

AI-driven changes in internet search behavior are impacting businesses like HubSpot, which lost significant web traffic due to AI's influence. HubSpot's Chief Marketing Officer, Kipp Bodnar, notes that AI tools, including large language models, are altering how users find information, reducing click-through rates by up to 70%. Companies are adapting by employing answer engine optimization (AEO) to improve visibility in AI-generated responses. This involves creating concise, relevant content that AI can easily extract. Businesses like Spice Kitchen and MKM Building Supplies are focusing on establishing authority and expertise to be cited by AI tools, which are increasingly preferred over traditional search engines like Google. AI visitors are proving more likely to convert into customers, highlighting the importance of adapting to this shift.

Meanwhile.

OpenAI is advocating for companies to experiment with a four-day work week as artificial intelligence becomes more prevalent in the workplace. The company suggests this as part of a broader "people-first" policy to help society adjust to the AI era, which promises both benefits and disruptions. OpenAI's proposals, primarily aimed at the United States, include creating more jobs in sectors like childcare, education, and healthcare. The company emphasizes that AI advancements could significantly reduce the time needed for certain tasks, potentially reshaping organizational operations and job opportunities. To prepare for this shift, OpenAI recommends that businesses consider enhancing worker benefits, such as maintaining pay levels during a four-day work week, increasing retirement contributions, and subsidizing healthcare and childcare. Professor Gina Neff from the University of Cambridge notes that while the concept of compensating workers for efficiency gains from technology isn't new, OpenAI's approach requires a significant shift in political and economic dynamics. This discussion arises amid concerns that AI could displace jobs, similar to the Industrial Revolution's impact. However, some experts argue that AI's effects on jobs and productivity may take longer to manifest than anticipated. OpenAI's suggestions echo those of other AI firms, highlighting the need for societal and economic adjustments to accommodate AI growth.

In other news.

Courtney Blackann reports that Catenya McHenry, chief marketing and communications officer at St. Stephen's Episcopal School, emphasizes the importance of using data to enhance storytelling in communications. McHenry's team at the school leverages engagement data to identify which stories resonate most with their audience, focusing on human interest stories that highlight student experiences. By reviewing data weekly, including open rates and engagement metrics, the team adjusts their content strategy to prioritize themes that consistently perform well, such as student achievements. This data-driven approach allows them to create content that not only engages their current audience but also attracts prospective families by showcasing student life through videos and social media posts. McHenry stresses the need to combine analytics with an understanding of human behavior to refine storytelling techniques. Her team tracks performance over various timeframes to demonstrate the impact of their strategies to leadership, showing how specific content decisions lead to measurable outcomes. This approach has resulted in steady growth on social media without additional spending, proving the value and efficiency of their communication efforts. McHenry will discuss these strategies further at Ragan’s Employee Communication Conference, highlighting the importance of aligning content strategy with organizational goals.

After that.

India Inc faces a challenging workforce environment with rising employee expectations and cautious AI adoption, according to Mercer’s Global Talent Trends survey. Only 30% of employees are very satisfied and plan to stay, while 28% are satisfied but intend to leave. Pay is the main reason for attrition, with 54% willing to switch jobs for better compensation. AI presents both opportunities and concerns, with 79% trusting their organizations to provide AI skills, yet 75% worry about increased surveillance. Companies are preparing for structural changes, with 76% planning to flatten hierarchies. Purpose-led work is a strong motivator for Indian employees, cited by 74%, compared to 42% globally. Organizations must redesign work to enhance performance and growth.

Finally.

The New York Times features letters from readers expressing concerns about the impact of artificial intelligence on writing and publishing. Margaret McGirr from Greenwich, Conn., compares AI-generated content to "canned music," emphasizing the importance of human emotion in writing. Bill Henderson of Pushcart Press notes the challenge of distinguishing AI-generated works from those crafted by humans, highlighting the potential for legal action against fraudulent submissions. Bruce Fagerstrom from Darien, Conn., shares his apprehension about competing with AI in fiction writing, stressing the value of human creativity and authenticity. These letters reflect a broader anxiety about AI's role in creative fields and its potential to overshadow human contributions.

In a separate letter, Harvey Olin from Parkland, Fla., responds to David French's column on American polarization. Olin agrees with French's assessment that the rise of divisive leaders like Donald Trump is symptomatic of a deeper civic crisis. He underscores the need to rebuild a civic culture that values disagreement without dehumanization and calls for a Congress that upholds its constitutional duties. Olin echoes French's warning about the dangers of a political climate that fosters division and erodes democratic norms. The letters collectively highlight concerns about technological and political challenges facing contemporary society, emphasizing the need for human-centered approaches and civic renewal.