An article in Scientific American by Archon Fung and Lawrence Lessig presents a parallel approach to my research on democracy and artificial intelligence. Citizens’ assemblies are a good way to fix this, by solving emotionally charged issues prior to an election. Even better, nested assemblies could replace much of the election process.
Sen. Josh Hawley asked OpenAI CEO Sam Altman this question in a May 16, 2023, U.S. Senate hearing on artificial intelligence. Altman replied that he was indeed concerned that some people might use language models to manipulate, persuade and engage in one-on-one interactions with voters.
Here’s the scenario Altman might have envisioned/had in mind: Imagine that soon, political technologists develop a machine called Clogger – a political campaign in a black box. Clogger relentlessly pursues just one objective: to maximize the chances that its candidate – the campaign that buys the services of Clogger Inc. – prevails in an election.
As a political scientist and a legal scholar who study the intersection of technology and democracy, we believe that something like Clogger could use automation to dramatically increase the scale and potentially the effectiveness of behavior manipulation and microtargeting techniques that political campaigns have used since the early 2000s. Just as advertisers use your browsing and social media history to individually target commercial and political ads now, Clogger would pay attention to you – and hundreds of millions of other voters – individually.
Adam Cronkright
Ali
Ahmed R. Teleb
avanderven
avergne
Alex Kovner
azakaras
Brett Hennig
campbellwallace
Simon Threlkeld
Nick Coccoma
codyhipskind
Common Lot Sortitionist
Conall Boyle
constitutionalism
davidschecter
graytmatter
gtridimas
H.J. Hofkirchner
John Burnheim
jgarry20151
Jochen Krattenmacher
johnedwardtaylor
johnhuang865
Jonathan Crock
keithsutherland
Jacob Richter
kkoreilly
lukewhito
SimonPek
Nathan El
Nicholas Gruen
Owen Shaffer
Patrick Chalmers
Paul Gölz
peterstone
Pierre-Etienne Vandamme
pronfeld
Emil
rcaze
richwbrown
Roger Knights
roslynfuller
Pierre-Etienne Vandamme
shaharavin
tbouricius
Ted Wachtel
Joshua Laferriere
theodorestathis
towardsautonomyblog
Wayne Liebman
Yoram Gat
Oliver Milne
Equality by lot 
Why, in principle, would citizens chosen by allotment be less open to individually-targeted manipulation? A more likely outcome would be the claim that AI would make democracy unnecessary as Mr Altman can operationalise the wisdom of crowds more accurately than a random sample. It would also appeal to those epistemic democrats who claim that the general will is an objective (rather than constructive) reality.
LikeLike
The programmable voter, AI edition.
LikeLike
The advantage of allotted juries over voters against this technology is that the juries can be presented with the evidence and engage in group deliberation largely in person, rather than online.
LikeLike
True, but who decides what is “evidence” and who gets to present it? These are political, rather than technocratic, issues. It would be nice to think that there was some impartial mechanism that could be called on, but Alex and I have chosen instead to seize the nettle, as that strikes us as the democratic route.
LikeLike
> The advantage of allotted juries over voters against this technology is that the juries can be presented with the evidence and engage in group deliberation largely in person, rather than online.
The primary advantage of the allotted over voters is that they can set their own agenda and options, which indeed requires deliberation (and indeed being physically present together makes a crucial difference in the quality of the deliberations).
Voters cannot deliberate – either in person or online – due to the huge numbers involved. They can only express their opinions on and preferences between options which were set by others. Even if those preferences are very well informed and considered, whoever sets the menu of options controls policy.
LikeLike
That’s clearly the case, the only caveat being that there is no good reason to believe that agenda setting by a tiny group of volunteers will match the agenda of the overwhelming majority of citizens disenfranchised by the lottery. This will not worry deliberative and epistemic theorists, but it should ring alarm bells for those who still believe in democracy.
LikeLike
Has there been any research comparing the quality of in-person and zoom deliberation? My anecdotal impression is that the latter can sometimes be better, and less prone to the groupthink created by social pressures.
LikeLike
These questions (how to structure and select balanced information and experts, whether to meet in person or online, etc.) are why I advocate a rotating but standing mini-public charged solely with making adjustments and improvements to the processes and design of policy juries. (I call this a Rules Council in my multi-body paper)
Some deliberative tasks are likely better done in person, while a final judgment (yes/no vote on adopting a law) might be done better remotely. Give and take deliberation CAN result in group-think, deference to high status members, etc. Jeremy Bentham advocated that only written arguments be used by legislators in order to avoid untoward psychological influences of skillful orators and focus attention on the logical merits of the arguments. Individual internal deliberation (as opposed to active give and take deliberation) also embraces the independence necessary to satisfy Condorcet’s “jury theorem” and to harness the sort of wisdom of crowds demonstrated by Galton’s agricultural fair steer weight guessing investigation.
As for selecting evidence, facts and witnesses, there are many options….One idea is to have all opposing interests assemble sets of facts and stipulate (as in a court trial) to which “facts” are universally agreed to, and which are disputed. But clearly the minipublic needs to be able to call in witnesses of their choosing as well.
In short there is not one design that works best for all functions. So, to optimize a sortition system, different bodies with different functions with specifically designed (and constantly refined) procedures is best.
LikeLike
Terry:> all opposing interests assemble sets of facts
That’s what Alex and I are proposing, the only difference being that the parties involved need to pass a minimum threshold to secure a place at the table. Note that the etymology of “party” is simply a partial interest.
>stipulate (as in a court trial) to which “facts” are universally agreed to, and which are disputed.
That’s a problematic (and unnecessary) task. Both democratic theory and the argumentative theory of reasoning would suggest that the jury (aided by the advocates) should be the body which separates the wheat from the chaff.
>different bodies with different functions
But why should all these bodies be appointed by sortition?
LikeLike