AI round-up 2025

This year can be called the year of artificial intelligence, as the tech is here to stay.

The AI landscape has shifted from simple chatbots to AI interfaces performing complex tasks. Here are key concepts that shaped the industry in 2025, according to MIT researchers.

Superintelligence
Tech giants like Meta and Microsoft pivoted their entire strategies toward being the first to lead in creating AI models with “superintelligence”—a level of AI that surpasses human capability.

Meta CEO Mark Zuckerberg created a superintelligence research lab, luring in top AI researchers with million-dollar salaries, and even slashed his much-touted metaverse funding to focus on AI.

Reasoning models
This year was defined by a massive push toward reasoning models like OpenAI’s o1 and DeepSeek’s R1 that can tackle complex problems step by step.

AI slop
The term “slop” became widely used by the public due to the flood of low-quality, AI-generated content on social media feeds.

Rise and fall of vibe coding
The vibe coding trend—where users could build apps or websites with simple text prompts on AI chatbots—rose to popularity in the summer and appeared to cool down soon, according to Barclays data.

GEO
Search Engine Optimization (SEO) is being replaced by Generative Engine Optimization (GEO). Businesses are now scrambling to ensure their content is cited by AI models rather than just appearing in a list of links.

Tools like Google’s AI Overviews and AI Mode now give users direct answers instead of traditional blue links. It hit news websites and publishers the most as search traffic to major outlets has declined.

Agentic AI
Companies like Salesforce touted their AI agents. CEO Marc Benioff confirmed in September that his company cut around 4,000 customer support jobs as more AI agents took up the roles.

Not just companies, people also increasingly use AI assistants for work and personal tasks. Retail analysts predicted that this holiday shopping season would see more customers using AI assistants to discover products.

Hyperscalers and backlash
As AI grew, so did the demand for massive data centers, referred to as hyperscalers. Projects like OpenAI, Oracle, and Softbank’s $500 billion Stargate facilities occupy several acres of land and employ thousands of people.

However, they face criticism over high energy consumption and its impact on local utility bills.

Overall, AI became more entrenched in society and is no longer just a trend.