AI trends

🧰 Emerging AI trends

2025:

1. Agentic AI
2. AI Governance Platforms
3. Disinformation Security
4. Post-Quantum Cryptography
5. Ambient Invisible Intelligence
6. Energy-Efficient Computing
7. Hybrid Computing
8. Spatial Computing
9. Polyfunctional Robots
10. Neurological Enhancement

See 2025 trends →

2024:

1. Generative AI
2. Overlap of AI and IoT
3. Augmented intelligence
4. Transparency
5. Composite AI
6. Algorithmic decision-making
7. No code tools
8. Cognitive analytics
9. Virtual assistants
10. Wearable devices
11. Data security and regulations
12. Natural language processing

See 2024 trends →

An appetite for AI knowledge

---

🔥 The State of AI

1. In-demand skills: ~50% of respondents are using reskilling to build AI talent
2. Performance focus: 2^x greater likelihood that companies who lead in AI have hired a machine learning (ML) engineer
3. Diversity challenge: 25% average share of racial or ethnic minorities developing AI solutions
4. Sust-AI-nable: 43% say their organizations are using AI to assist in sustainability efforts
5. The bottom line: ~25% of respondents says that ≥5% of their organizations' EBIT is attributable to AI

---

🔍 AI Search Tools

This is a new section this year, and respondents' top choice for AI search tools is ChatGPT: 83% of respondents have used it in the past year. This is above and beyond the second choice of Bing AI with 20% having used it. The hype around emerging AI search technology has room to grow while the ChatGPT competitors grow their user base; only four tools had 10% or higher selection for those that want to try it in the next year.

Developers want to keep using ChatGPT for their AI Search. Other tools they want to use are Phind and WolframAlpha.

---

🔨 AI Developer Tools

In addition to asking about search tools beginning this year, we also asked about AI developer tools. GitHub Copilot is the overall pick for most used AI developer tool with 55% of respondents using it this past year, quadrupling the second top pick of Tabnine at 13%. Those learning are using Tabnine a little more (18%) and Copilot a little less (45%), possibly due to costs associated with Copilot.

Developers want to continue using GitHub Copilot and, in a flip, we see more developers overall who want to try it over the next year than those currently using Copilot.

---

🐞 Why companies struggle with AI?

Unfortunately for all of us, this is a highly accurate depiction of #MLOps. Read "Machine Learning Operations (MLOps): Overview, Definition, and Architecture" by Dominik Kreuzberger.