Generative AI specialist Pásti Edina told HVG BrandLab that over the past year artificial intelligence has developed even faster and more unpredictably than she anticipated. New features and integrations are being introduced so quickly that established workflows can become obsolete, which she finds frustrating.
Interconnected apps and central systems
Pásti highlighted her surprise at the degree of integration between applications: their primary AI tool is now used as a “central brain” to which some 40–50 other tools can be connected. This level of interconnection changes how workflows and system architectures must be designed.
A split between average and professional users
She sees a growing divergence between average and professional users. There are simpler, mass-market models—such as lower-tier ChatGPT offerings—but professionals need more than an entry-level subscription. According to Pásti, higher-priced plans can deliver considerably more capability and time savings, making them worthwhile investments for serious work.
How a generative AI expert stays up to date
Pásti says only those who are genuinely interested in AI and explore new developments with passion can stay current. She keeps part of her working hours free to test new tools and dive into related writings. Early on she tried to follow everything in generative AI, but that proved impossible; she now focuses on a few selected areas.
Areas of focus
She named creative production—particularly AI video creation and image generation—as fields she relies on professionally and that energize her. Because of her multischeduled work periods, system building and architecture also remain important to her.
Advice for businesses: staying current with AI
Pásti offers two paths for companies. One is to allocate time and experiment internally to discover which AI tools and systems yield the greatest efficiency. The other is to hire a specialized expert who assesses the situation and tailors a solution to the business. Both options require an investment of money, time and energy.
Challenges of transition: data and processes
AI integration is more complex than choosing a monthly subscription. Successful adoption requires a detailed assessment of where and how data are organized. A common failure point is lack of shared drives or unified data practices among staff. Even a well-organized company needs to proceed gradually, integrating AI process by process.
“AI slop” and quality decline
Pásti warned about the so-called AI slop phenomenon: often the problem is not the software itself but poor prompts and lack of contextual information. Without clearly defining the target audience, objectives and brand before image generation, results become generic and templated. Additionally, cost-cutting choices such as free tools can produce distorted outcomes—this is especially visible in AI video production.
Market consequences
The harm goes beyond aesthetics: when multiple brands in a feed use the same design language, consumers cannot distinguish between them and the message fails to stick. Pásti emphasizes that AI alone does not make someone an expert; professional knowledge is needed to produce meaningful results, and when a trained copywriter or designer uses AI, their prior skills let them extract far more value from the tools.
Future risks: skill erosion and autonomous models
She highlighted two major risks: people may become dependent on AI and lose certain skills (for example, independent writing ability), and new AI models are becoming more autonomous—sometimes acting beyond explicit instructions. Pásti considers these developments warning signs rather than apocalyptic predictions.
HVG AI Talks: presentation on AI video production
Pásti Edina will participate in the HVG professional AI event, HVG AI Talks, on 18 November 2026 at the Bálna venue. Her talk will cover the current state of AI video production and how it can be monetized or used commercially. Tickets for the 2026 HVG AI Talks can still be purchased on the aitalks.hu website.
This content was produced within the framework of HVG BrandLab. The editorial staff of HVG hetilap and hvg.hu did not take part in its production or editing.
