The pile of applications is tall, and you're tempted to let a tool boil each one down to three lines. That can save effort, but it can also hide the person you'd most want to meet. Here's how to use summaries and still catch what they miss.
What goes wrong with summaries
A summary is someone else's choice of what matters. When the "someone" is a tool, the choice may not match what matters for your job.
- Details get left out. A short, odd career path or a skill described in unusual words may not make the summary.
- Details get added. A tool can state something the CV never said. Only the original shows you.
- Gaps get treated as faults. The UK government guide says "A sifting tool may also consider gaps in employment in a way that disproportionately affects parents" and others with care duties, disabled people and neurodivergent candidates.1
- Guesses about the person. The UK data regulator found tools "inferring characteristics, including gender and ethnicity, from a candidate" name.2
None of this rules out summaries. They just can't be the only thing you read, because a tool that skims fast will also skim past the quiet, strong candidate.
Step 1: Write your criteria before you look
List what the job needs, in your own words, before you read any application. Keep it to skills, experience and availability that matter for the work. Use the same list for everyone. That keeps you honest about why one person looks stronger than another.
The EEOC states that "An employer may not base hiring decisions on stereotypes and assumptions" about protected characteristics.3 A written list gives you something concrete to hold yourself to.
Step 2: Read the originals
Read each application yourself. It takes longer. It's also the only way to be sure you haven't missed someone worth meeting. If the pile is too big to read at once, read in rounds, and don't skip any round for anyone.
Step 3: Use summaries as a map
Once you've read an application, a summary can help you find a detail again, compare people against your list or prepare questions. Treat it as a note that might be wrong.
If you do paste anything, remove names and contact details first. Privacy is the reason, and the applicant-privacy page covers it. Better still, paste only your own notes about each person.
Step 4: Check each summary against the original
Open both side by side and check:
- Does every skill in the summary appear in the CV?
- Is anything in the CV missing that matches your list?
- Has the summary judged the person, with words like "weak" or "strong"? Delete those. Your criteria and your own reading make that call.
- Does it mention a gap or a change of direction as a negative? Ask the person about it in the interview instead.
Step 5: Ask what a tool might have learned
The UK government guide says "screening tools are often trained using historical data".1 That creates a risk the system "might inherit bias from past recruitment practices".1 A general chat tool isn't a screening product. The same worry fits, though: it may favour familiar patterns over unusual good candidates.
That's why the shortlist stays yours. If someone's application looks strong on paper to you but a summary played it down, trust your reading of the original.
What the regulators point to
The ICO says AI is used to "summarise CVs and score applicants".2 Scoring and ranking people is the step where mistakes become decisions. Another page explains why that step stays with a person.
Reference list
- Responsible AI in recruitment, Department for Science, Innovation and Technology (UK)
- ICO intervention into AI recruitment tools leads to better data protection for job seekers, Information Commissioner's Office (UK)
- Prohibited Employment Policies/Practices, U.S. Equal Employment Opportunity Commission