AI is expensive. Human intelligence is more devious. When they meet, it is match made in heaven.<\/p>\n\n\n\n
Amazon wanted to hasten AI adoption. They gamified the use of AI agents by employees. The leaderboard, called \u201cKiroRank\u201d, ranked employee use AI. Employees were \u2018caught\u2019 tokenmaxxing, using AI to perform random, and useless tasks. Employees responded exactly as rational humans do when subjected to a badly designed metric.<\/p>\n\n\n\n
Amazon eventually shut the leaderboard down.<\/p>\n\n\n\n
The story is funny, not is haha funny, but in the way organizational absurdities often are. It reveals something deeper about this phase of the AI boom: much of what is being sold as earth shattering is actually hype. i started my career in dot boom , and in Han Solo immortal words, “Didn’t we just leave this party?”. <\/p>\n\n\n\n
This is not a story about workers misunderstanding artificial intelligence. It is a story about workers understanding incentives perfectly. There is an old principle in economics called Goodhart\u2019s Law: when a measure becomes a target, it ceases to be a good measure.<\/p>\n\n\n\n
I first learned this in school, though nobody called it Goodhart\u2019s Law then. Social service was a subject. We got marks for how much \u201cservice\u201d we had done. We were encouraged to go to hospitals and get our work attested. The nurses were busy. They signed the forms. Often, no service had happened. The moment service became marks, it stopped being service and became paperwork.. One days teachers caught on.<\/p>\n\n\n\n
I lived through the era of ISO certification. Every organisation was applying for one. Whether they needed or not, Corporate\/ bureaucracy genes are remarkably stable. The ritual changes. The paperwork changes. The dashboards become more sophisticated. The underlying behaviour does not.<\/p>\n\n\n\n
The moment \u201cAI usage\u201d became a metric, it stopped measuring meaningful adoption. it become performance \u2013 dekho.. \u201csee how I am using AI\u201d.<\/p>\n\n\n\n
Workers did what workers have always done inside metric-heavy systems. They optimized for what was visible. For all the breathless rhetoric about machine intelligence transforming everything, its future is already visible. It is going to be co-opted by bureaucracy.<\/p>\n\n\n\n
This is not unique to Amazon.<\/p>\n\n\n\n
It is the same behaviour that gave us schools teaching to standardized tests instead of teaching children. The children, meanwhile, learned exactly what they were being taught: how to pass a test.<\/p>\n\n\n\n
It is what happens when call centers optimize for shorter call times rather than solving problems. Press 1 to be told your call is important.<\/p>\n\n\n\n
It is why social media platforms optimized for engagement and industrialized outrage. Nobody set out to radicalize lonely young men. The algorithm simply noticed that anger travels further than happiness.<\/p>\n\n\n\n
Now the same logic is being applied to AI. Managers need proof that they are \u201cdoing AI.\u201d Boards need numbers to show shareholders. Middle management needs following the boss orders. . So institutions create measurable proxies for transformation. Prompt counts. Token usage. Adoption rates. Output skyrockets, but outcome become wonky.<\/p>\n\n\n\n
\u201cI maxxed claude usage\u201d
\u201cdoing what?<\/p>\n\n\n\n
Amazon\u2019s internal AI leaderboard is a reminder that when technology is adopted without consciousness, change becomes a item dance. Hyped up:, sexy, but story can continue without it.<\/p>\n\n\n\n
The question is no longer: How do we prove we are using AI?<\/p>\n\n\n\n
It is:<\/p>\n\n\n\n
What is this helping us become?<\/p>\n\n\n\n
A brief note:<\/em><\/p>\n\n\n\n I am recovering from aphasia following a stroke. AI assists me in shaping language as my speech and grammar continue to heal.<\/p>\n\n\n\n The thinking, analysis and perspective remain entirely my own.<\/p>\n\n\n\n <\/p>\n","protected":false},"excerpt":{"rendered":" Amazon\u2019s abandoned AI leaderboard is not really a story about artificial intelligence. It is a story about incentives.<\/p>\n Employees responded exactly as rational humans do when a metric becomes a target: they optimized for the metric. <\/p>\n From school social-service certificates to ISO audits, from engagement algorithms to AI adoption dashboards, organizations keep repeating the same mistake. The ritual changes. The paperwork changes. The dashboards become more sophisticated. The underlying behaviour does not.<\/p>\n The question is no longer whether we are using AI.<\/p>\n The question is whether AI is helping us create value, or merely helping us perform compliance more efficiently. \ud83d\ude08<\/p>\n","protected":false},"author":1,"featured_media":8723,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"Artificial intelligence meets old-fashioned babudom.\n\nGuess who won?","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":false,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"jetpack_post_was_ever_published":false},"categories":[1976,15],"tags":[120,2008],"class_list":["post-8716","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-artificial-intelligence","category-economics","tag-ai","tag-goodhart-law"],"yoast_head":"\n