In recent months, Human Resources departments around the world have experienced an unprecedented phenomenon: a veritable avalanche of job applications written by artificial intelligence.
Tools like ChatGPT have democratized access to polished, affordable and keyword-laden writing, which has skyrocketed the number of applications reaching companies.
This digital tsunami threatens to collapse traditional selection systems and poses important ethical and operational challenges.
The figures that shake Human Resources
According to data from LinkedIn, the platform processes around 11,000 applications per minute, a figure that has grown by 45% compared to the previous year.
This growing volume is forcing recruiters to turn to their own AI solutions to filter out those applications before they end up sitting in a mailbox unchecked.
The result is a dynamic in which, on the one hand, candidates rely on algorithms to highlight their skills and, on the other, companies hire algorithms that decide who deserves an interview.
Although at first glance it may seem that automation speeds up the process, the truth is that many of these applications are indistinguishable from each other.
With text structured in a uniform way and loaded with predefined “keywords”. These are terms that the applicant management system (ATS) recognizes as valuable.
In the end, recruiters find themselves filtering through millions of generic and precisely optimized documents. AIs generate profiles thatmeet formal requirements but lack individuality and authenticity.
Latent discrimination in algorithms
When the writing of a resume is outsourced to an AI, not only the form is delegated, but also the substance. Language models learn from huge amounts of texts on the Internet, where historical prejudices of gender, race and social classes nest.
For example, recent studies have shown that classification systems secretly favor male profiles when analyzing candidates for positions traditionally dominated by men.
Similarly, video analysis programs in AI interviews have shown biases against non-native accents or against people with certain facial expressions.
These biases are especially worrying because they operate under a layer of “mathematical objectivity” that recruiters assume is infallible. When a system rejects a candidate, it is rarely questioned whether the reason is based on unfair data.
The consequence: those already vulnerable groups tend to receive fewer opportunities, deepening the diversity gap in organizations.
Impact on diversity and corporate culture
The homogenization of curricula also erodes diversity of thought. If all applicants present skills and achievements in a similar way, companies lose the possibility of discovering unique profiles that, despite not adapting their CV 100% in ATS format, could provide freshness and innovation.
This uniformity runs the risk of translating into teams that are more similar to each other and less creative, just the opposite of what many companies defend in their corporate values.
The AI arms race
Far from abating, the AI storm in selection processes is intensifying. As AI-generated resume detection systems become more refined, so do text generators.
This dynamic is reminiscent of an arms race: each advance by one side is countered by an advance by the other.
Emerging companies like Jobright.ai offer services that go beyond writing a resume: they automatically complete the entire application process, personalize cover letters and optimize LinkedIn profiles. Its users claim to have doubled their interview opportunities.
In response, HR departments are incorporating AI “signal” detectors (repetition of syntactic structures, excessive use of synonyms or standard formats) to separate those candidates that lack a real human touch.
Regulation and institutional responses
Faced with this panorama, legislators are beginning to raise their voices. In New York, Local Law 144 requires independent audits of AI systems used in selection processes, in order to identify and correct possible biases.
Illinois has regulated the use of automated video interviews since 2023, requiring companies to inform candidates of their use and obtain their express consent.
In the European Union, the recent Algorithmic Transparency Directive provides for sanctions for those who deploy decision engines without offering avenues for appeal or human review.
As usual, regulations still lag behind technological advances. Many SMEs and startups do not have the resources to implement in-depth audits or hire algorithmic ethics experts.
Consequently, legal recommendations are sometimes limited to “best practices” that few organizations consistently apply.
Keys to adapt to the new panorama
Faced with this silent revolution, both candidates and companies must redefine their strategies:
- Complementary human review: Incorporate at least one manual evaluation stage, where a professional contrasts the authenticity and coherence of the profile.
- Training in ethical AI: Train HR teams. HH. in the new risks associated with algorithms, so that they know how to detect biases and anomalies.
- Transparency with candidates: Clearly inform when and how automated systems are used, and offer appeal channels.
- Encouraging originality: Encourage applicants to include concrete examples of projects, measurable achievements, and personal anecdotes that stand out beyond the keywords.
These measures will not completely eliminate the challenges, but they will help balance the balance between efficiency and equity. Artificial intelligence will continue to transform the way we recruit talent, but the human factor must continue to be the definitive filter.
This post is also available in: