- CORPGRO Newspoint
- Posts
- COMPLY OR AI
What’s the Story?
In January press reports indicated that JP Morgan intends to use an in-house AI tool called Proxy IQ, which will analyse governance compliance at 3,000 US listed companies. Reliance on this tool means JPM will no longer retain the Proxy Voting Agency ISS.
Glass Lewis, the other Proxy giant has flagged in recent times too that their approach to governance analysis will become more bespoke and granular, and that this extra workload and detail will be made possible by use of AI empowered tools.
The US Republican agenda in recent times has also criticised the weight given by Proxys to ESG factors, particularly climate, as business unfriendly and a drag on economic growth.
Proxy voting agencies have also been criticised as holding excessive power over governance and institutional investors and outcomes.
Why Does it Matter?
The AI self-service initiative by JP Morgan may roll-out worldwide, and be adopted by other leading investors such as Blackrock etc.
Comply or explain has long been the Corporate Governance mantra. Comply is the expected default but explain is accepted where the explanation is detailed and relevant to the business. These governance norms flow deeply into the world of executive compensation as a strongly contentious area of institutional investor opinion and voting.
In practice this works well for the global giants, but less well for the vast majority of listed companies, where “Comply or Else” seems the main approach from the proxy agencies.
The advent of AI in churning through the vast amount of detail which is inevitably involved in comply or explain should be welcome. The box ticking compliance approach has been criticised strongly as a “no-brain” exercise adding no value. If this is displaced by AI with more speed and accuracy, what is not to like?
Well, some companies already fear to “explain” given the strains imposed by the explain route. To comply is an easier and safer choice even if it is sub-optimal.
A big unknown is how the governance community will configure around these AI tools to produce votes in favour or against in the new environment.
Alongside that concern are the AI issues of bias, hallucination, psychosis and slush.
AI should be fast and accurate, but when it is not, the flaws are deep and hard to spot. AI has generated Ovid poems to explain comments on Roman history, in Latin obviously, and in Ovid’s style, but these poems were not written by Ovid. Similarly, the UK West Midlands police used AI to help justify a ban on a visiting Tel Aviv football team, but part of their evidence was behaviours at a match that never occurred. AI invented both the fixture and the claimed poor behaviours.
On the plus side, if AI frees up expert human time to consider fully the explain non-comply choices that would be a good outcome. But only if the displaced person hours are then focussed on the tricky judgemental issues of explain. On the minus side though, AI may just decrease governance cost and headcount to the detriment of professional judgement.
It is too early to say which will be seen, likely a bit of both - but as cost is such a clear and measurable driver, the danger is that this cost cut impact will predominate by far.
One global head of Corporate Governance explained some time back that their desire for in person meetings with Remco Chairs was essential so that:
“I can feel if the handshake is clammy, or the eyes shifty and if they speak with ease on the structure and rationale.”
You do not get that from AI.
The AI LLMs will have been trained on the vast amount of corporate governance thinking and publications developed over many decades and in many jurisdictions. How the governance community manages the ongoing development of the received body of good practice guidance to companies and voting institutions is unknown. The role of governance bodies such as the ICGN is likely to be enhanced.
Newspoint view
AI is inevitable, but how we use the tech and our control protocols around reliance and results are not.
In the first Industrial Revolution the Luddites had a point, yes economic growth was delivered, but alongside a sharp rise in industrial accidents and miserable working conditions. Better organised labour would have eased the process; indeed, organised labour arose as a result.
Technological change is now natural and should be welcomed; but the focus should be on the needed new controls and how the benefits are shared, including broad based employee share ownership.
Technology has already changed our world beyond the imagination. Our planet supports 8.3bn people. Thomas Malthus, a Victorian economist, predicted famine as global population grew. He did not foresee the green revolution of the 1960s nor the untold advances in technology and productivity we now take for granted.
So, technology is generally a good thing with some disruptive aspects, AI will evolve despite some future bumps. How that plays out for governance compliance is as yet unknown, but this is an issue that can be shaped.
CORPGRO Helps Companies With:
Please feel free to email or call:
Damian Carnell - [email protected] +44 (0) 7989 337118
VA Bec Bostock - [email protected]
Please share this CORPGRO information with your board or your colleagues.