Grantmakers urged to eschew grants ‘beauty contests’

Posted on 19 Nov 2026

By Kathy Richardson, Executive Directors, SmartyGrants

Share this article

Magnifying glass transparency investigation

Public sector grant managers should resist judging grant applications by the quality of the prose and instead focus on public value, a gathering of Australian Commonwealth grantmakers has heard.

Nick Sellars
Nick Sellars

Integrity expert, Churchill Fellow and renowned “grants nerd” Nick Sellars encouraged grantmakers to resist engaging in a “beauty contest” that selected the best-written applications, and focus instead on the plausibility of the applicant’s theory of change.

Grantmakers should ensure there was a clear line of sight from the application to their policy objectives, Mr Sellars told the SmartyGrants meeting in Canberra this month.

Fellowship report
Tap on the cover to access Sellars' full report.

Mr Sellars outlined some common traps in grantmaking, including seeking to “score the unscorable” (apply scores to highly subjective criteria), failing to identify and tackle entrenched biases, and failing to apply the “can they do it?” test (judge the applicant’s ability to achieve their objectives).

He said Commonwealth grant funding was increasing despite budget pressures, with a growing emphasis on “impact” rather than “stimulus”. 

To ensure proper use of public money, grants needed to be effective, efficient, economical and ethical, he said. 

Mr Sellars said there was not enough appreciation of the knowledge, skills and abilities needed to administer grants well. 

“Great grants and great grantmakers bring policy ideas to life,” he said. “You need to be agile, focused, and have a learning mindset.”

“You have to treat AI like a team member, like an intern – a team member you have to train for many months before they become useful.”
Paola Oliva-Altamirano, chief data scientist, SmartyGrants

The impact of AI on applications

https://files.ourcommunity.com.au/general/SG/Articles-SG/OlivaAltamirano_Paola.jpg

SmartyGrants chief data scientist Paola Oliva-Altamirano told the meeting that research published by the Research on Research Institute (RoRI) had found that the number of applications to 12 major research funders across six countries had increased by an average of 57% between 2022 and 2025.

Dr Oliva-Altamirano said the quality of applications had improved significantly as well, with artificial intelligence levelling the playing field and helping to eliminate the grants “beauty contest” that Mr Sellars had referred to earlier. 

This meant grantmakers had a harder job sifting and sorting applications, and they too were turning to AI to handle the increased load, she said.

Dr Oliva-Altamirano outlined how the use of AI could save time but could also produce a high error rate, so it needed to be handled carefully.

“In SmartyGrants we think of AI as more like a thinking partner than something that does your job,” she said.

“You have to treat AI like a team member, like an intern – a team member you have to train for many months before they become useful.” 

Dr Oliva-Altamirano said grantmakers needed to recruit carefully (pick the right AI tool), induct it as you would an intern (give it proper context, rules and boundaries – harnesses), subject it to probation (including testing it on work you’ve already completed), provide proper supervision (spot-check its work, set an error budget), give detailed feedback, and review its work – then promote it or let it go. 

She urged grantmakers to check three things when recruiting an AI tool:

  1. Data safety – where does your data go, who can see it, and is it used to train the model?
  2. Compliance – does it comply with laws and rules? Can it show you the evidence?
  3. Ethical framework – which ethical framework has been used (if any), and how is it reflected in the product itself?

Dr Oliva-Altamirano outlined how SmartyGrants applied a “dignity lens” (developed by transdisciplinary academic Lorenn Ruster) to the design of its AI tools, with close consideration of four factors:

  • How to keep a human in the loop
  • How to train the user so they can make good use of the tool
  • How to maintain visibility and traceability 
  • How to keep data safe. 

These rules were shaping the developing of an upcoming SmartyGrants eligibility assessment tool, she said.

Dr Oliva-Altamirano concluded by providing some practical governance tips for grantmakers considering using AI:

  1. Make it contestable – can an applicant challenge an AI-informed outcome? Who reviews it?
  2. Iron out the kinks by testing your tool on past rounds before it touches a live round.
  3. Spot check a sample of your data set and track the misses.
  4. Tell applicants how you want them to use AI, and how you are using it too.
  5. Plan your exit – consider a fallback in case the model you have chosen disappears or falls over. 

More news