AI in Marketing

Stan Berteloot, the curator of AI in Marketing, is grateful to be invited to an important conversation: how a Haitian university should govern students' use of AI without sacrificing its core purpose. This is not a debate about banning a tool. Students already use AI. The real question is whether higher education remains a place where minds are formed, not just outputs produced. If AI can draft, summarize, and argue on command, then the value of the university shifts toward what AI cannot provide: judgment, verification, intellectual courage, and the ability to defend ideas in public. Around the world, universities are converging on a pragmatic approach: faculty autonomy within clear guardrails, strict rules for data and privacy, and a move away from unreliable detection toward transparency, citation, and reflection. Some now formalize “process over product” by asking students to submit prompt trails, revisions, and reasoning, not just final answers. Others are redesigning assessment into two lanes: AI-free formats that verify knowledge, such as in-class writing and oral exams, and AI-enabled work that evaluates responsible, critical use. In Haiti, this moment is even sharper. Language, culture, and access gaps can distort what AI returns and who benefits. Done well, an AI policy becomes more than a regulation. It becomes a renewed commitment to the university's cognitive mission and a clear statement of what we mean by “human” in a world shaped by AI.

That said.

Emmanuel Macron defended the European Union's AI regulations at the AI Impact summit in Delhi, countering US criticism and emphasizing the need for child protection against digital abuse. He highlighted the misuse of AI, referencing Elon Musk's Grok chatbot, which generated sexualized images of children. Macron advocated for a collaborative approach to make the internet safer, proposing a ban on social networks for children under 15 in France. António Guterres supported these views, stressing that AI should not be controlled by a few entities. Narendra Modi emphasized AI's potential and India's role as a major player, advocating for open-source technology. The summit also saw tensions among tech leaders, with calls for international AI regulation and oversight.

On a different note.

Google has launched Gemini 3.1 Pro, an update to its AI model, introducing a three-tier reasoning system that allows users to adjust the model's computational effort. This update marks a shift in Google's release strategy, moving from full-version launches to more frequent incremental upgrades. The new system offers low, medium, and high reasoning levels, enabling a single model to handle tasks ranging from quick responses to complex problem-solving. This flexibility could simplify enterprise AI deployment by eliminating the need for multiple specialized models. Gemini 3.1 Pro has shown significant improvements in benchmarks, particularly in reasoning and agentic capabilities. It scored 77.1% on the ARC-AGI-2 benchmark, more than doubling its predecessor's score and outperforming competitors like Anthropic's Sonnet 4.6 and OpenAI's GPT-5.2. The model also excelled in other benchmarks, such as Humanity's Last Exam and GPQA Diamond, showcasing its enhanced reasoning and knowledge evaluation skills. The decision to label this update as 3.1 rather than a preview suggests substantial improvements, while the "point one" designation indicates an evolutionary step. The model is available in preview through various Google platforms, with further advancements expected before a full general availability launch. This release may prompt competitors to respond quickly in the rapidly evolving AI landscape.

After that.

India's largest manufacturers are collaborating with global industrial software leaders Cadence, Siemens, and Synopsys to enhance AI-driven design and manufacturing using NVIDIA's AI infrastructure, CUDA-X, and Omniverse libraries. The country is investing $134 billion in expanding manufacturing capabilities across sectors like construction, automotive, renewable energy, and robotics. Reliance New Energy is integrating Siemens' digital twin technology with NVIDIA Omniverse for advanced gigafactory design. Addverb Technologies and Hero MotoCorp are also leveraging NVIDIA's solutions for robotics and product development. Companies like Havells India and Larsen & Toubro Semiconductor are using Synopsys and Cadence tools for improved simulation and chip design. Tata Consultancy Services and Wipro PARI are advancing industrial automation with NVIDIA's AI platforms, enhancing safety and operational efficiency.

Meanwhile.

TechCrunch reports that as AI companies grow, debates continue about AI's impact on employment. While AI can automate tasks, some experts believe it will create new jobs, with human roles evolving rather than disappearing. David Shim, CEO of Read AI, compares AI's role to using maps in cars, emphasizing that humans will remain central in decision-making. He acknowledges AI's potential to replace certain jobs, such as in advertising, but notes the need for human oversight in automation. Abdullah Asiri, founder of Lucidya, argues that AI will replace tasks but not entire roles, allowing employees to take on new responsibilities like supervision and business development. AI tools, like Read AI's meeting notetakers, free up time for more strategic tasks. Both Read AI and Lucidya aim to maintain lean teams while enhancing productivity through AI. Read AI's sales tool, for instance, uses CRM data to predict deal outcomes, reportedly approving deals worth $200 million. Asiri stresses the importance of hiring AI-literate employees, as the skill is still developing. Customer perception of AI is improving, with users valuing issue resolution over whether a human or AI handles their queries. Asiri highlights that customers prioritize fast and accurate solutions, regardless of the method.

Next up.

Plato, a Berlin-based startup, has secured $14.5 million in seed funding to integrate generative AI into the wholesale distribution sector, which handles a significant portion of global goods but often relies on outdated systems. The funding round was led by Atomico, with participation from Cherry Ventures, Discovery Ventures, and D11Z. Plato aims to embed AI into existing ERP systems, transforming historical sales data into automated actions such as generating quotes and identifying risks. The company, founded by Benedikt Nolte, addresses inefficiencies he experienced in his family's distribution business. With several large distributors already on board, Plato plans to expand into procurement and customer service automation, targeting new European markets and eventually the U.S., marking a shift in AI investment focus.

In other news.

ATM jackpotting attacks are increasingly prevalent, with hackers stealing millions, according to a recent FBI bulletin. This criminal activity, which involves forcing ATMs to dispense cash without affecting customer accounts, has evolved from theoretical demonstrations to a significant threat. In 2025, over 700 attacks resulted in at least $20 million in stolen cash. Hackers gain physical access to ATMs using generic keys and employ digital methods, such as malware like Ploutus, which targets the Windows operating system of ATMs. Ploutus exploits XFS software, enabling hackers to control ATMs and dispense cash rapidly. These attacks are challenging to detect until after the cash is withdrawn, posing a serious security risk to financial institutions and ATM manufacturers.

Finally.

AI systems are increasingly taking over critical security decisions within corporate networks, as highlighted by PYMNTS.com. Offensive AI agents are probing for vulnerabilities, while defensive AI systems autonomously detect and mitigate threats, often before human analysts can respond. This shift is driven by the growing complexity and speed of AI-driven cyberattacks, which can adapt and generate unique attack instances that challenge traditional detection methods. The World Economic Forum notes that 87% of organizations see AI-related vulnerabilities as a rising risk. Gartner predicts that by 2027, 17% of cyberattacks will use generative AI. Companies like Cogent Security are investing in autonomous remediation, using AI to prioritize and execute vulnerability fixes, significantly reducing response times. However, the effectiveness of these systems depends on high-quality data, as poor data can lead to false positives or missed threats. Additionally, attackers are deploying fraudulent AI assistants to exploit user trust, emphasizing the need for robust AI security measures.