# The Good the Bad and the AI: Will We Let the Algorithm Choose for Us | Maria Dimitrova | TEDxPlovdiv

Source: https://www.youtube.com/watch?v=r8UngtzYJEo
Recap page: https://rapidrecap.app/video/r8UngtzYJEo
Generated: 2025-11-11T23:39:17.225+00:00

---
## Quick Overview

The speaker argues that relying solely on AI for recruitment, like the CV sorting done by Jobs.bg, is risky because the AI lacks human intuition and might mistakenly reject potentially excellent candidates who don't fit the programmed criteria, as demonstrated by the fact that 80% of qualified candidates interviewed were not the perfect match for the initial job description.

**Key Points:**
- The speaker introduced "Geri," a highly enthusiastic recruitment team member who joined the company three years prior, bringing excellent knowledge of HR and human relations.
- The company received 73 job postings, resulting in 28,311 CVs, 18,303 tests, 933 hours of work, 2,044 hours of analysis, 133 working days, and 292 working days for processing in 2024.
- Over two years, 81.6% of approved CVs in Q3 2025 passed the initial screening, but only 26.7% of those who applied for an interview were considered qualified based on the initial criteria.
- The speaker shares an anecdote about a candidate whose CV was poorly written (male, warehouse worker role) but possessed the necessary intuition, highlighting the risk of relying only on algorithms.
- The AI system, which the company uses to filter candidates, is programmed with specific criteria that may overlook candidates whose inner qualities or intuition do not align with the initial, narrow job description.
- If candidates are filtered based only on the initial job description, companies risk losing potentially excellent hires who might only be revealed after a short interview or through human assessment.
- The presentation emphasizes that while AI is useful, it should not entirely replace human judgment, as the best candidates might not fit the algorithm's rigid profile.

![Screenshot at 00:05: Speaker Maria Dimitrova standing on stage in front of large red and white TEDx Plovdiv letters, beginning her presentation on AI in recruitment.](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-00-05.png)

**Context:** Maria Dimitrova speaks at TEDxPlovdiv about the challenges and risks associated with over-reliance on Artificial Intelligence (AI) in the recruitment process, using concrete data from her company's experience since integrating an AI filtering system from Jobs.bg. She contrasts the quantitative results of AI screening with the qualitative, human aspects of candidate evaluation, emphasizing the importance of intuition and nuanced assessment over strict algorithmic matching.

## Detailed Analysis

Maria Dimitrova begins by introducing Geri, a key member of her recruitment team who joined three years ago, bringing valuable expertise in HR and human relations. She presents impressive 2024 statistics showing the high volume of applications received: 73 job postings led to 28,311 CVs reviewed, 18,303 tests administered, requiring 933 hours of work and 292 working days of processing. She projects significant improvement in screening metrics into 2025, with accepted CVs rising to 81.6% by Q3 2025, although only 26.7% of those who applied for an interview actually met the initial criteria. Dimitrova then uses an analogy about buying a coffee machine: sometimes the machine is great, but sometimes you just want a simple cup of water, illustrating that AI-driven processes might filter out good candidates whose qualities don't match the initial, narrow parameters. She shares a story where the AI filtered out an excellent candidate because the CV was poorly written, despite the candidate possessing the necessary intuition. She concludes that while AI is a powerful tool for optimization, relying solely on it risks eliminating truly valuable hires whose potential is only revealed through human interaction or intuition, stressing the need for human judgment alongside the technology.

### Recruitment Data (2024/2025 Projections)

- 73 job postings
- 28,311 CVs reviewed
- 18,303 tests passed
- 81.6% CV acceptance rate projected for Q3 2025

### The AI Filtering Risk

- AI filters based on rigid criteria, potentially overlooking excellent candidates whose intuition or unstated qualities don't match the initial description

### The Coffee Machine Analogy

- Sometimes a simple cup of water is preferred over a complex coffee machine, symbolizing the need for simpler, intuitive human assessment versus complex AI filtering

### Case Study of a Rejected Candidate

- An excellent candidate with strong intuition was filtered out because their CV was poorly written, illustrating the limitations of automated screening

### Conclusion on AI Use

- AI is a useful tool, but human intuition is necessary; relying solely on AI can lead to losing great talent, as demonstrated by the 26.7% interview rate discrepancy.

### Final Advice

- Do not let the algorithm choose for us; ensure that rigorous, human-centric criteria are applied during the selection process.

![Screenshot at 00:00: Opening shot showing the TEDx Plovdiv stage setup with large red and white letters.](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-00-00.png)
![Screenshot at 00:10: Close-up of Maria Dimitrova speaking into a headset microphone on stage.](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-00-10.png)
![Screenshot at 01:32: Slide titled "Резултати" \(Results\) displaying recruitment statistics for 2024 and projected for 2025 quarters.](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-01-32.png)
![Screenshot at 02:21: Slide showing data highlighting the gap between approved CVs \(81.6% in Q3 2025\) and those interviewed \(26.7% in Q3 2025\).](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-02-21.png)
![Screenshot at 03:57: Speaker gesturing emphatically while discussing the risks of relying on AI filtering.](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-03-57.png)
![Screenshot at 06:06: Speaker standing on stage, demonstrating the scale of applications received.](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-06-06.png)
![Screenshot at 07:36: Speaker pointing out the low percentage of candidates who pass the initial screening based on algorithmic criteria.](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-07-36.png)
![Screenshot at 11:07: Slide displaying a sample job description for a "Warehouse Worker, Regular Working Time" highlighting sections likely reviewed by AI.](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-11-07.png)
![Screenshot at 16:04: Slide showing the final slide with the text "HR Human & Robots" indicating the central theme of the talk.](https://ss.rapidrecap.app/screens/r8UngtzYJEo/00-16-04.png)